Speed Learning...(IJMECS-V8-N7-3)

P3701 | NVIDIA JETSON AGX ORIN SERIESDATASHEETDiscover the most powerful AI computer for energy-efficient autonomous machines.NVIDIA ® Jetson AGX Orin ™ series modules deliver up to 275 TOPS of AI performance with power configurable between 15W and 60W. This gives you more than 8X theperformance of Jetson AGX Xavier ™ in the same compact form-factor for robotics and other autonomous machine use cases.These system-on-modules support multiple concurrent AI application pipelines with an NVIDIA Ampere architecture GPU, next-generation deep learning and visionaccelerators, high-speed IO, and fast memory bandwidth. Now, you can develop solutions using your largest and most complex AI models to solve problems such as natural language understanding, 3D perception, and multi-sensor fusion.Jetson runs the NVIDIA AI software stack, and use case-specific application frameworks are available, including NVIDIA Isaac ™ for robotics, DeepStream for vision AI, and Riva for conversational AI. You can also save significant time with NVIDIA Omniverse ™ Replicator for synthetic data generation (SDG), and with NVIDIA TAO Toolkit for fine-tuning pretrained AI models from the NGC ™ catalog.Jetson ecosystem partners offer additional AI and system software, developer tools, and custom software development. They can also help with cameras and other sensors, as well as carrier boards and design services for your product.Jetson Orin modules are unmatched in performance and efficiency for robots and other autonomous machines, and they give you the flexibility to create the next generation of AI solutions with the latest NVIDIA GPU technology. Together with the world-standard NVIDIA AI software stack and an ecosystem of services and products, your road to market has never been faster.Jetson AGX Orin 32GB>1792-core NVIDIA Ampere architecture GPU with 56 tensor cores >2x NVDLA v2.0>8-core Arm® Cortex®-A78AE v8.2 64-bit CPU>32GB 256-bit LPDDR5 >64GB eMMC 5.1 >PVA v2.0Power>Voltage input 5V, 7V-20V >Module Power: 15W - 40WKey FeaturesJetson AGX Orin 64GB>2048-core NVIDIA Ampere architecture GPU with 64 tensor cores >2x NVDLA v2.0>12-core Arm® Cortex®-A78AE v8.2 64-bit CPU>64GB 256-bit LPDDR5 >64GB eMMC 5.1 >PVA v2.0Power>Voltage input 5V, 7V-20V >Module Power: 15W - 60WNVIDIA JETSON AGX ORIN SERIES MODULES TECHNICAL SPECIFICATIONS* Virtual channel-related camera information for Jetson AGX Orin is not final and subject to change.Refer to the Software Features section of the latest NVIDIA Jetson Linux Developer Guide for a list of supportedfeatures.。

前向神经网络学习速率的自适应算法(英文)
刘巧歌;付梦印;邓志红
【期刊名称】《系统仿真学报》
【年(卷),期】2006(18)3
【摘要】学习速率是控制神经网络学习过程的一个重要参数,影响神经网络的稳定性和快速性。

提出了一种能够满足实时性要求的神经网络学习速率的自适应算法,并证明了在该学习速率下,神经网络的学习过程是Lyapunov意义稳定的。

该方法通过为神经网络的输出增加一个输出修正量来补偿多个未知因素对学习误差的影响,从而构造使学习误差快速收敛到零的学习速率自适应算法。

通过对神经网络在线逼近一个非线性对象的过程进行仿真,结果证明了该方法的有效性。

【总页数】4页(P698-700)
【关键词】学习速率;学习误差;神经网络;BP算法
【作者】刘巧歌;付梦印;邓志红
【作者单位】北京理工大学信息科学技术学院自动控制系
【正文语种】中文
【中图分类】TP18
【相关文献】
1.一种前向神经网络自适应学习算法 [J], 高书江;李成峰
2.一种学习速率自适应的可编程片上学习BP神经网络电路系统的设计 [J], 卢纯;石秉学;陈卢
3.一种新的前向神经网络自适应学习算法 [J], 余庆梅;曲蔚
4.基于硬限幅功能函数的前向神经网络的分类学习算法(英文) [J], 张兆宁;孙雅明;毛鹏
5.基于学习速率自适应的脉冲神经网络快速学习方(英文) [J], 方慧娟;罗继亮;王飞因版权原因，仅展示原文概要，查看原文内容请购买。

I.J. Modern Education and Computer Science, 2022, 4, 57-66Published Online August 2022 in MECS (/)DOI: 10.5815/ijmecs.2022.04.05Solution for Using FEMM in Electrostatic Problems with Discrete Distribution Electric ChargeMihaela OsaciPolitehnica University of Timisoara, Revolutiei no.5, Hunedoara, RomaniaCorina Daniela CunțanPolitehnica University of Timisoara, Revolutiei no.5, Hunedoara, RomaniaIoan BaciuPolitehnica University of Timisoara, Revolutiei no.5, Hunedoara, RomaniaReceived: 26 January 2022; Accepted: 19 April 2022; Published: 08 August 2022Abstract: Finite Element Method Magnetics (FEMM) is an open source software package for solving electromagnetic problems based on the finite element method. The application can numerically solve linear electrostatic problems and magnetostatic 2D problems, respectively low frequency magnetic, linear harmonic and nonlinear. FEMM is a product much used in science and engineering that, in the last 15 years, has begun to be used more and more in the academic environment. Despite the fact that FEMM can be used to solve complex problems in science and engineering, electrostatic FEMM cannot work directly with discrete electric charge distributions, that is, point electric charge. This work presents a FEMM model for simulating point electric charge that can be used in case of electrostatic problems with discrete charge distributions. The numerical solution for the electrostatic field is compared with the analytical solution. This model can be used in the case of an assembly of point electric charges with axial symmetry.Index Terms:Teaching tools, modalities of teaching, computer-assisted training, finite element method, numerical modeling in electromagnetism.1.IntroductionIn the age of digital technology, computer-assisted training is one of the basic teaching tools of modern education [1, 2, 3]. Electromagnetic field theory is one of the most difficult courses studied by students of the specialization of electrical engineering and other related specializations. Since few problems of electromagnetic field theory have analytical solutions, the use of software packages for modeling and simulation are extremely necessary, make learning in this field much more efficient and improve the individual and team work of students [4, 5]. Since numerical modeling in electromagnetism has become a tool extensively used by engineers, scientists and researchers [7], in electrical engineering programs, lately, there is a growing emphasis on teaching design-oriented courses [6]. This category mainly includes courses that are based on the application of techniques of numerical modeling of the electromagnetic field in solving specific problems in the technique.In the modern teaching of numerical methods for the electromagnetic field very useful are the software applications with graphical interface that allow the design, implementation, processing and post-processing of an electromagnetic field problem. For university use, and generally in the educational environment, such an application should be easy to use, offer programming support in scripting language, The ability to import and export files and further support in engineering design activity. Among applications with a graphic interface for the numerical solutions of differential equations with partial derivatives by the finite element method here: ANSYS [8, 9], Comsole Multiphysics [10, 11] and FEMM (Finite Element Method Magnetics) [12, 13, 14, 15].FEMM is an open source analysis software package with finite elements for electrostatic, magneto-static and low frequency electromagnetism problems. It is a simple to use, good quality, accurate and low cost freeware product usedin both science and engineering. Besides electromagnetism, it helps to solve complex problems in other areas such as materials science, industry, medicine, physics, robotics, astronomy and space engineering [16, 17, 18, 19, 20, 21, 22, 23, 24].For some 15 years, there have been a number of concerns aimed at exploring the ability of FEMM to meet the needs of electromagnetism teaching in higher education and beyond. A number of quantitative and qualitative studies and illustrative examples are intended to support the usefulness of this package of programs in the educational environment [13, 14, 15].Although FEMM is a widely used software in electromagnetism, has no support for electrostatic problems with discrete distribution of electric charge.In the specialized literature are offered two types of solutions for the implementation of FEMM of the load with discrete distribution [22]: the introduction of the discrete load through the properties of the node and the introduction of the discrete load as load of the inner armature of very small radius of a cylindrical capacitor implemented planar. The first solution lead to distort the shape of the electric field lines. The second solution is forced because in the planar implementation the depth parameter must be introduced, which means the length of the capacitor .Our work presents a solution for using FEMM software in issues with discreet electric charge distribution, much closer to reality, which can help improve the learning experience in electromagnetism at both university and college levels.Our work reinforces the idea that FEMM can also be used in classical areas [24], such as physics, where more emphasis is placed on the analytical aspect of problem solving, with electric and magnetic fields often seen as abstract notions, difficult to understand by students and students.2. FEMM Model for Point Electric ChargeFEMM is a 2D finite element modeling tool. In general, to solve a problem with the FEMM, the problem design is defined and created in the appropriate pre-processor and then one of the FEA solvers available for that type of problem is used [25]. Finally, the results can be analyzed in the postprocessor. The application interface is easy to use and allows to define the problem adding suitable objects, symbols and materials. DXF files with problem design or expanded metadata can also be imported. The application supports LUA scripting using, which allows to schedule specific actions. a. Electric charge point - analytical modelFrom the electrostatic (Coulomb’s theorem) it is known that the electric field created in the air, for example by a point electric charge q at the distance r of the electric charge is in size:24qE r πε= (1)And the electrical potential at a point at a distance R from the load creating the field is:4qV r πε= (2)For example, using a Matlab script it is easy to calculate the electric field and potential at a given distance from an electric charge that creates this field-fig.1.Fig. 1. Matlab script for the electric charge-analytical solutionRunning this script for the distance of 1cm and a positive electric charge of 1C gives the results from Fig.2 and Fig.3.Fig.2. Analytical results - electric chargeFig. 3. Graphical representations of the analytical solution: E=f(r) și V=f(r)b.Electric point charge - FEMM modelOnce a planar type problem has been defined, with units of measure for cm-fig length. 4, the first idea of implementing a point-of-form electric charge in FEMM starts with the point property definition, with “Add property” -fig.5. In the “Nodal Property” dialog fill in “Name” and select “Point Charge Density Property”. Due to the parameter depth - Fig. 4, for total charge of 1C, at “Point Charge Density, C/m” must be entered 100 (100C/m 1cm=1C).Fig. 4. Defining the problem with the natural idea of implementationFig. 5. Electric charge simulation through “Nodal Property ” Attach the property to the creeat point and define the boundary condition (Dirichlet condition, i.e. the electrical potential is null on the border).We will use the “Asymptotic Boundary Condition” method (as described in the Appendix to the FEMM manual [25]) to mimic an unbounded geometry. -fig.6. The results are shown in Fig.7. It is noticed that the results of the numerical simulation are far from the analytical results, which means that the natural idea presented just now is far from reality.Fig.6. Stages in the implementation of electric charge through “Nodal Property ”Fig. 7. Results of the implementation of electric charge through “Nodal Property ” The solution we propose starts from the observation that if in the expression of the electrical capacity of a spherical capacitor we make the outer radius to tend to infinity, we obtain the electrical capacity of a metal sphere charged withthe electric charge q - table 1, whose electrical potential is determined from q C V =,4r q q V C rπεε==, dV Edr =-,where 204r dV q E dr rπεε=-= as shown by the relation (1). Thus, the point electric charge can be simulated by a very small conductive metal sphere with a given charge, by defining an Axisymmetric Problem type-Fig. 8.Table 1. Electrical capacity and electric field created by a metal sphere charged with electric chargeFig.8. Defining the problem in the proposed solutionThe geometry of the problem is then constructed by placing the coordinate points (r,z) =(0,0.1) and (r,z) =(0,-0.1) that join by an arc to which the Conductors property is assigned-Fig.9.Fig.9. Defining the conductors property "Sim_sarc" and its association to the model arc An open border condition predefined by FEMM is used, with the Air material setting in the field of solutions -fig.10. The FEM network is generated and the solver is launched into execution and the result is obtained from fig.11.Fig. 10. The model of point electric chargeFig. 11. Numerical results3.Results and DiscussionFig. 11. shows that the numerical result is very close to the analytical result. For a better concordance can be refined the mesh of finished elements - fig. 12.Fig. 12. Numerical results after refining the finished item networkFig.13. shows the graphical representations of the numerical solution. Comparing these representations with the graphical representations of the analytical solution, a very good concordance is observed.Fig. 13. Graphical representations of the numerical solution: E=f(r) și V=f(r)The comparison between the analytical solution and the numerical solution can be better done by reading Table 2.Table 2. Comparison between the analytical solution and the numerical solution proposedType of solutionMesh size 0.1 Mesh size 0.01 E(V/m) V(V/m) E(V/m) V(V/m) Analytical solution9⋅1013 9⋅1011 9⋅1013 9⋅1011 Numerical solution 8.73707⋅1013 8.96668⋅1011 9.01085⋅1013 8.98723⋅1011For a better comparison between the analytical solution and the numerical solution, can be calculated the relative error in accord with relation:analytical solution numerical solution(%)100analytical solution E -=⋅ (3)In case of mesh size 0.1 the relative error is 2.92% and in case of refined mesh it obtained the relative error 0.12%. This analysis shows a very good concordance between the analytical values and the numerical values.3.1 Use of FEMM model for point electric charge for a set of point electric chargesThe model shown in the first subparagraph may be used in the case of a system of point electric charges if the system shows axial symmetry. For example, let's consider the example from din fig.14a, with q 1=10μC, q 2=5μC, r 1=1cm and r 2=2cm.a. b. c.Fig. 14. System of 2 point charges 3.2 Analytical solutionThe programming of the analytical solution was done in Matlab-fig 15, the solution being shown in fig. 16.Fig. 15. Matlab script for the analytical solution -system of electric chargesFig. 16. Analytical results - system of electric charges3.3Numerical solutionTo implement the numerical solution we choose the axis of symmetry to pass through the 2 charges, fig. 14b. Rotating fig. 14b. until the axis of symmetry reaches the upright position, fig. 14 c, the geometry of the problem to be solved using the point electric charge model shown in the previous paragraph is obtained. The numerical solution to the problem is shown in Fig. 17.Comparing the analytical solution of the problem shown in Fig. 16 with the numerical solution represented in Fig. 17, a very good concordance is observed.Fig. 17. Numerical results - electric charge system4.ConclusionsThe paper presents a solution for FEMM implementation of a point electric charge and the use of the proposed model in solving an electrostatic field problem for the case of a system of point electric charges. The model can be successfully used in electrostatics courses at both the university and college levels for a better understanding of the issue. The advantage of using the FEMM software package in the didactic activity is very high, given the fact that it is free and open source, easy to use and offers students the possibility to understand abstract concepts such as those of electric and magnetic fields and offers them skills in the design of electromagnetic devices, useful skills in their careers as future engineers.References[1]Y. Kong and I. Xie, “Professional Courses for Computer Engineering Education”, I. J. Modern Education and ComputerScience, 1, pp.1-8, 2010.[2] A. Herala, A. Knutas, E. Vanhalan and J. Kasurinen, “Experiences from Video Lectures in Software Engineering Education”,International Journal of Modern Education and Computer Science, Vol.9, No.5, 2017.[3]M. Mladenović, M. Rosić and S. Mladenović, “Comparing Elementary Students' Programming Success based on ProgrammingEnvironment”, International Journal of Modern Education and Computer Science, Vol. 8, No. 8, 2016.[4]K. K. Krutikov and V. V. Rozhkov, “Features of Electrical and Magnetic Skin Effect Modeling from AlternatingElectromagnetic Fields in FEMM”, Russian Electrical Engineering, Vol. 91, pp. 781–785, 2020[5]K.G. Brandisky, K.P. Stanchev, I.I. lacheva, R.D. Stancheva, S.K. Petrakieva, S.D. Terzieva, V.M. Mladenov, “Computer-Aided Education in Theoretical Electrical Engineering at the Technical University of Sofia: Part II”, in EUROCON 2005 - The International Conference on "Computer as a Tool", 2005[6] D. Kacprzak, P. Surdacki, H.D. Stryczewska, B. Guillemin, “Magnetic modelling projects in university courses - NewZealand and polish examples”, in IET 7th International Conference on Computation in Electromagnetics (CEM 2008), pp. 66 –67, 2008[7] D. Kacprzak, “Implementation of Finite Element Method Modelling Tools in Education Programs”, in6th InternationalConference on Computational Electromagnetics, 2006[8]“ANSYS”, [Online]. Available: https:///. [Accessed 2 March 2022][9]T. Stolarski, Y. Nakasone, S. Yoshimoto, Engineering Analysis with ANSYS Software, second edition, Elsevier PublishingHouse, 2018[10]W. B. J. Zimmerman, “Comsol Multiphysics and the Basics of Numerical Analysis”, Series on Stability, Vibration and Controlof Systems, Series A: Vol. 18 -Multiphysics Modeling with Finite Element Methods, pp. 27-63, 2006[11]“COMSOL”, [Online]. Available: https:///. [Accessed 2 March 2022][12]“FEMM-homepage”, [Online]. Available: https:///wiki/HomePage. [Accessed 2 March 2022][13]K. B. Baltzis, “The finite element method magnetics (FEMM) freeware package: May it serve as an educational tool inteaching electromagnetics?”, Education and Information Technologies, vol. 15(1), pp.19-36, 2010[14]K. B. Baltzis, “On the usage and potential applications of the finite element method magnetics (FEMM) package in theteaching of electromagnetics in higher education”, in 8th International Conference on Computer Based Learning in Science (CBLIS 2007), 2007[15]R. Crozier and M. Mueller, “A new Matlab and Octave interface to a popular magnetics finite element code”, in XXIIInternational Conference on Electrical Machines (ICEM), 2016[16]L. Hao, L. Xue, F. Huang, G. Cai, W. Qi, M. Zhang, Q. Han, Z. Wang, J. Lin, “A Microfluidic Biosensor Based on MagneticNanoparticle Separation, Quantum Dots Labeling and MnO2 Nanoflower Amplification for Rapid and Sensitive Detection of Salmonella Typhimurium”, Micromachines, Vol.11(3), pp. 281, 2020[17]T. A. Elmasri, M. A. Elmasri and E.S. Abdulhafid, “Finite Element Analysis of Free Energy Permanent Magnet Motor UsingSolidworks and Finite Element Method Magnetics (FEMM) Software”, Journal of Marine Sciences & Environmental Technologies, Vol. 5, Issue 2 , 2019[18]G. Priyandoko, P. Suwandono, N.R. Ismail, W.M. Utomo and S. Ubaidillah, “Development of Vibration IsolatorMagnetorheological Elastomer Based”, Journal of Physics: Conference Series, Vol. 1908, The 1st International Conference on Innovation and Application of Science and Technology (ICIASTECH 2019) 2-3 October 2019, Malang, Indonesia, 2021 [19]I. F. Lopes, D. C. Coelho, E. V. A. Bojorge, L. R. A. de Oliveira, “Underwater Wireless Power Transfer With High Toleranceto Misalignments”, Brazilian Power Electronics Conference (COBEP), 2021[20]M. Ya. Marusina and A. A. Silaev, “Improving the Efficiency of Mechatronic Systems Based on OptimizationPrinciples”, International Conference on Quality Management, Transport and Information Security, Information Technologies (IT&QM&IS), 2021[21] D. Leonardis, C. Loconsole and A. Frisoli, “A passive and scalable magnetic mechanism for braille cursor, an innovativerefreshable braille display”, Meccanica, vol. 55, pp.1639–1653, 2020[22]https:///en/document/read/7522512/simulation-of-electric-and-magnetic-fields-using-femm-fh-aachen-[23]G.Anghel, I.A. Anghel, M.D. Calin, E. Helerea, “Magnetics Tutorial. Modelling Permanent Magnets Using theElectromagnetic Software FEMM 4.2”, The International Scientific Conference eLearning and Software for Education;Bucharest "Carol I" National Defence University, Vol. 4, pp. 267-274, 2018[24]M. Boulé, “The role of Finite Element Method software in the teaching of electromagnetics”, Fourth InterdisciplinaryEngineering Design Education Conference, 2014[25]“FEMM-manual”, [Online]. Available: https:///Archives/doc/manual42.pdf. [Accessed 2 March 2022]Authors’ ProfilesMihaela Osaci, Lecturer, Polytechnica University of Timisoara, Engineering Faculty of Hunedoara /ElectricalEngineering & Industrial Informatics Department. Main activities and responsibilities: didactic and scientificresearch activities. Technical skills and competences: Physics, Electromagnetism, Nanomagnetism, Modeling andSimulation.Corina Daniela Cunţan, Lecturer, Polytechnic University of Timisoara, Engineering Faculty of Hunedoara/Electrical Engineering & Industrial Informatics Department. Main activities and responsibilities: didactic andscientific research activities. Technical skills and competences: Electronics Automatic management processes andApplications of electronics in industrial systems.Ioan Baciu, Lecturer, Polytechnic University of Timisoara, Engineering Faculty of Hunedoara /ElectricalEngineering & Industrial Informatics Department. Main activities and responsibilities: didactic and scientificresearch activities. Technical skills and competences: Electronics and Quality of electrical energy.How to cite this paper: Mihaela Osaci, Corina Daniela Cunțan, Ioan Baciu, "Solution for Using FEMM in Electrostatic Problems with Discrete Distribution Electric Charge", International Journal of Modern Education and Computer Science(IJMECS), Vol.14, No.4, pp. 57-66, 2022.DOI: 10.5815/ijmecs.2022.04.05。

I.J. Modern Education and Computer Science, 2016, 9, 28-34Published Online September 2016 in MECS (/)DOI: 10.5815/ijmecs.2016.09.04Survey on Adverse Effect of Sophisticated Integrated Development Environments onBeginning Programmers’ SkillfulnessAlaba T. OwoseniDepartment of Computer Science, Interlink Polytechnic, Ijebu Jesa, Osun State, 233114, NigeriaEmail: atimothyowoseni@S. A. AkanjiDepartment of Mathematics and Statistics, Interlink Polytechnic, Ijebu Jesa, Osun State, 233114, NigeriaEmail: deerious1@Abstract—Integrated development environment as a software system that aids programmers in developing software applications quickly and effectively has been perceived to also serve as an inappropriate tool for beginning programmers when it is specially developed with some complex features. This complexity in features as perceived leaves the programmers with a double role of studying complexity found in the environment and the semantics with syntaxes of the concerned programming language. This paper categorizes few of the available integrated development environments based on program building tools that are integrated in them and also considers an experimental survey on their adverse effects on novice programmers by sampling programmers’ opinions using closed ended questionnaires. The population was randomly selected from some tertiary institutions in Nigeria. The opinions were statistically analyzed using chi square and based on the analysis, beginning programmers learning strengths are found greatly influenced by the type of integrated development environment used.Index Terms—IDE, integrated development environment, compiler, intelliSense, interpreter, effect of integrated development environment on programmers.I.I NTRODUCTIONComputer programming is a difficult and much challenges-demanding task executed by students [1][2] and even by any person. The challenges involved in programming might not be a point of discussion to an experienced programmer due to some motivational factors and also the level of skillfulness but these are always points to pound on when there is a population of novice under study. There have been various views on various challenges been faced by new programmers at their earliest years of programming. Some of these challenges are, learning language syntaxes, gaining access to computer systems or networks, learning language semantic structure, learning other program constructs such as comments, control structures, data types and so on, learning sub-programs, designing a program to solve a task, debugging, lack of competent tutors, lack of technical textbooks, and the complexities involved in integrated development environments. Many researchers before now had considered some of these challenges and the level of their effects on new programmers but, little or no work had been done on scientific confirmation of adverse effect of sophisticated development environments on learning strength of the new programmers. Therefore, this paper provides a survey on the adverse effect of sophisticated development environments on novice programmers.Integrated development environment (IDE) is a software system that provides comprehensive facilities to computer programmers for software development. During software development, there are many program building tools that a programmer uses in developing software applications. These program building tools are all integrated into an environment that is referred to as integrated development environment.Some IDEs are graphically enriched while some are not. Based on the number of program building tools that are integrated together to form an environment, an IDE may be categorized as sophisticated or non-sophisticated. Sophisticated development environment is an IDE that integrates many of the program building tools (debugger, compiler, editor and so on) and its features can only be easily understood by an experienced user (programmer) who is keen in using it for software development. Some of the integrated development environments in this category include; Microsoft Visual Studio, Netbeans, and so on. It is noted that majority of the graphical-based integrated development environments are sophisticated and support more than one language. However, the programming skills of some experienced programmers are heavily enhanced by these sophisticated IDEs while to the beginners, they serve as tools that do not enhance good programming skills due to their complexities. These complexities have always leave new programmers in a state where a single role of learning semantic and syntactic structures of programming language is doubledwith the cost of comprehending the complexities found by the IDEs to use.Non-sophisticated integrated development environment integrates few of the program building tools and its features can be easily understood by a user more so a beginner who wants to use it for software development.II.L ITERATURE R EVIEWA.Program Building ToolsThe integrated development environment as a system is made up of some program building tools seen as its components and among these are:1)Source Code EditorSource code editor is a text editor program designed specifically for editing source codes of computer programs by programmers [3]. Source code on the other hand is a file of computer instructions written in a particular programming language. Examples of source code editors are notepad, JEdit, notepad++ among others. Editors have auto-complete feature that automatically completes keywords or reserved words and also syntax highlighting feature that highlights reserved words or bugs in different colors as configured by the user.2)CompilerA compiler is a system software that translates a computer program written in a particular language (usually high level language) into an equivalent program in another language (machine language). Translation of programs here, is always done in its entirety before execution.3)InterpeterInterpreter is a software system that translates a computer program written in a particular high level language into an equivalent program in another language while the translation is done line by line or statement by statement.4)Documentation ToolThis a program building tool that is used for documenting the source codes. It uses to make source code more understandable and clearer to programmer for proper debugging and for future referencing most especially in this generation of code reuse.5)IntelliSenseIntelliSense is an integrated development environment feature that helps in automatic generation of code in the code editor. It helps to reduce time spent in typing code elements by programmers and helps to prevent the introduction of typographical errors in code.6)Build Automation ToolA tool that helps in scripting or automating a wide variety of compilation tasks, packing tasks, documentation tasks and development tasks that software developers do frequently[4].B.Overview of the Benefits of integrated DevelopmentEnvironmentThe benefits of integrated development environment as contained in [5] are:1)Maximization of programmer’s p roductivity with less effortIDE provides some features that assist programmers in maximizing their productivities. Little input or effort is expected from the programmers and this little input yields maximized products with the help of the features contained in the IDE.2)Reduction in software development timeWith the available program building tools, development time becomes shorter. With this tools, many of the needed codes are auto-generated thereby reducing the lines of code to be edited manually by the programmer.3)Enforcement of project or company standards Standards may be enforced if the IDE offers predefined templates or blueprints or if code libraries are shared between different team members working on the same project.4)Reduction in development stressThe stress passed through by programmers when developing complex software applications may be reduced with the help of IDEs.C.Related WorkAn exploratory study was carried out in [2] by computing department staff of National University of Samoa to investigate the most common errors students made in Java programming classes. There was an analysis of program code from undergraduate Java programming classes and results of analysis were used to form recommendation to inform courses’ development and improve teaching practices.In [6], a work on the review of literature relating to the psychological or educational study of programming for the purpose of identifying various problems experienced by novices was carried out. The problems include issues relating to basic program design, algorithm complexity in certain language features, fragility of novice knowledge and others but there seems to be little or no consideration on the complexity in IDEs. The researchers were later able to make some speculative observations and note possible topics for future work.Survey on difficulties faced by students in learning programming were considered in [10]. These difficulties involved the IDEs that were used by the programmers under study. The result of the research according to their responses showed that IDEs have great influence on the learning curve of programming students. However, the research does not specify the type of IDE and the class ofinfluence it has on programming students. More so, the population under study in the research was specific to the concerned population.As researched in [7], modern integrated development environments have tools that make recommendations and automate common tasks, such as refactoring, auto-completions, and error corrections. However, these tools present little or no information about the consequences of the recommended changes. Having to compute the consequences always puts an extra burden on the developers. But, the researchers developed a technique that could reduce the burdens encountered by the developers (who are believed to be experienced developers). This techniques informs developers of the consequences of any code transformation.[8] conducted a one-on-one interviews with 22 freshmen who were taking their first Java programming courses with the objective of investigating not only which programming concepts or constructs most students had difficulties with, but why they found them difficult. The focus was on fundamental object oriented concepts and Java programming constructs such as, classes vs. objects, static data members vs. constant data members, constructors, access modifiers, syntax for method calls parameter passing, method overloading, inheritance, polymorphism, foreach vs. for loops, and the use of standard Java libraries. Students’ misconceptions and missing conceptions in each of these concepts/constructs were described in detail in the research.A research [9] addressed the problem of IDE interface complexity by providing a single window graphical user interface. The main goal of the research was to create a usable graphical interface design for IDE without tool views with the aim that it will reduce the number of open tool views functionality into the text editor. However, there was only a partial implementation of this single-window design in KDevelop IDE and there was a need for future work on suability testing of single window interface to find whether usability problems are solved or at least reduced compared to traditional IDE graphic user interface.In [10], a research investigated and analyzed the problems faced by computer programming students at the University of Tabuk with two main objectives (finding out whether the students at the University of Tabuk faced problems in computer programming similar to the ones faced by the students in different corners of the world as reported in the literature and studying the impact of sociocultural and environmental factors on learning computer programming skills by the students of the University. The researchers designed questionnaires (that contain questions pertaining to educational facilities such as curriculum and teaching materials, lab equipment and class rooms’ environment, and the adequacy and proficiency of the professors and teaching assistants) to sample opinions and results from analysis of the questionnaires provided insight into the environmental and socio-cultural effects and the difficulties experienced while learning and teaching programming.Collaborative learning and its potential positive effect on the learning outcomes of programming students was investigated in [11] while [12] described an investigation into the nature of the academic problems that face novice programming students. The later research analyzed the results of a survey given to students enrolled in an introductory programming unit across three campuses at Monash University in 2007. The survey focused on student perceptions of the relative difficulty in understanding and implementing both low level-programming concepts, such as syntax and variables, and high level concepts, such as OOP principles and efficient program design. An analysis of the approximately 150 responses in the study indicated that a significant percentage of students experienced difficulties in high-level concepts.To this point in time, it appears that little work has been done to directly consider the effect of IDEs most especially those with complex facilities on new programming students.III.M ATERIALS AND M ETHODA.Integrated Development Environment Classification In this paper, the IDEs are classified into two based on the number of program building tools integrated into them. The two classes are non-sophisticated and sophisticated development environments.1)Non-sophisticated development environmentIt is an IDE that integrates few of the program building tools and its features can be easily understood by a new programmer who wants to use it for software development. An understanding of the technical knowhow of it requires little or no effort from the programmers who want to use it. Examples of these IDEs are:a)Turbo Pascal IDETurbo Pascal IDE is an IDE that is capable of running on some operating systems among which are windows based, DOS, Macintosh [13]. It integrates tools like text editor, compiler, linker and few other tools. It needs little or no training from new programmers before its usage. b)Qbasic IDEIt is also a simple to use IDE that does not require extra training before a new programmer can begin exploring its features and used for maximizing productivity. Few of the program building tools are integrated in it.2)Sophisticated Development EnvironmentIDE in this class integrates many of the available program building tools and its features and usage can only be comprehensive to experienced programmers who are keen in using it for software development. It usage demands for extra training because without training, many of its features cannot be maximally utilized. Some IDEs in this class are:a)ActivatdeState KomodoActivateState Komodo was developed by ActiveState [15]. It greatly supports Ruby, javaScript, XUL, Perl, Python, XHTML, CSS3 and Tcl. It is an IDE that is capable of running on some platforms such as Windows, Mac and Linux. It contains some features like auto-completion, syntax coloring as configured by users and source code control integration with subversion, Bazaar, Git, Mercurial, CVS and Perforce.b)NetbeansNetbeans is a multi-language IDE for developing primarily with Java, but also with some other languages, in particular, PHP, C/C++, and HTML, Ruby and so on. This IDE is written in java and can run on many operating system that are available in the market such as Windows, Linux, Solaris. Since it is a java application, it can run on any platform that supports java virtual machine [14]. This IDE, integrates many program building tools (code editor, intelliSense, and code refactoring tool, designer for designing GUI applications, web designer, class designer and database schema designer etc.) and its enhanced use needs training on how to use the available tools found in the environment. It is provides tools for developing applications for the three editions of java that is, standard, enterprise and micro editions.c)Microsoft Visual StudioThis is an IDE developed by Microsoft Corporation [15]. It is used for developing both console and graphic based applications, Windows forms applications, web applications and web services applications like as we have in Netbeans. It includes program building tools like code editor, intelliSense, and code refactoring tool, designer for designing GUI applications, web designer, class designer and database schema designer [16]. It is multi-language integrated development environment that provides support for some languages among which are C, C++, VB. NET, C# and F# [15].d)EclipseEclipse is a multi-language IDE developed by free and open source software community and it is written in java [9]. It runs on some operating systems such as Linux, Mac OS X, Solaris and Windows. It supports development in languages like Ada, C, C++, COBOL, FORTRAN, Haskell, Perl, PHP, Python, R, Ruby, Scala, Clojure, Groovy and Scheme [17].e)Oracle JDeveloperOracle JDeveloper is an integrated development environment developed by Oracle Corporation. It is written in java language and it is a freeware type IDE [10]. The features provided by JDeveloper helps to support developments of software applications in some programming languages such as Java, XML, PL/SQL, JavaScript, SQL and PHP among others [18]. With JDeveloper, Oracle has aimed to simplify application development by focusing on providing a visual and declarative approach to application development in addition to building and advanced coding environment [18]. It integrates with the Oracle Application Development Framework, an end-to-end Java EE based framework that furthers simplifies application development for programmers [18].3)Advantages of Sophisticated IDE on Experienced ProgrammersThe advantages of sophisticated IDE on experienced programmers include:∙Its effectiveness in developing complex software projects∙Its involvement in the reduction of software development time∙Its maximization of programmer’s productivity with little effort∙Its involvement in stress reduction during software development∙Its appropriateness as the right choice for an experienced programmer who has vast knowledgewith its use4)Advantages of Non-sophisticated IDEThe following are few of the advantages of the non-sophisticated IDE:∙Its easiness with technical knowhow since few tools are integrated.∙Its right choice for a beginner∙Its power of computer resources’ consumption in terms of memory, processors’ cycle and so on.5)Disadvantages of non-sophisticated IDETo an experienced programmer, the following are the disadvantages of non-sophisticated environment:∙Its less effectiveness for complex software projects ∙Its limitation in operation since few features are implemented in it.B.MaterialsDuring the course of studying the adverse effects of sophisticated IDEs on beginning programmers, we designed four hundred and twenty closed-ended questionnaires that contained questions as represented in table 1. These questionnaires were used in obtaining data from their primary sources (students and lecturers in programming related disciplines).C.Methods1)Data CollectionData collection in this paper has been done through questionnaire. The population for study was carefully selected and it comprised of computer science and computer engineering undergraduates and lecturers of some tertiary institutions in Nigeria. These institutionsare Federal University of Technology, Akure (FUTA), Ladoke Akintola University of Technology, Ogbomoso (LAUTECH), Usmanu Danfodiyo University, Sokoto (UDUS), Interlink Polytechnic, Ijebu Jesa (IPI), Osun State College of Technology, Esa Oke and Federal Polytechnic, Ede (FEDPOEDE). The population of our study were of various academic levels who have in one form or the other made use of IDEs that fall into the categories of IDEs under study.2)Data Presentation and Analysisa)Data PresentationThree hundred and eighty questionnaires were returned and this data as collected from the population were presented using simple pie chats as shown in fig 1. and fig 2. The collected data has been analyzed using a statistical tool called chi square as illustrated shortly.Fig.1. Presentation of returned questionnairesTable 1. Results of responses as contained in questionnaires.b)Data AnalysisThe levels of significance (α) in this analysis are 1% and 5% while the hypothesis are:Hypothesis TestedH0: Sophisticated IDEs do affect learning strength of new programming students.H1: Sophisticated IDEs do not affect learning strength of new programming students.E= (380/10) =38;Fig.2. Responses as collected through questionnairesX2 = ∑ ((O-E) 2/E) = 4.320247From the table 2 of the Chi-square distribution,D.f = X2 α (n-1) = X2 0.01, 9 = 21.67D.f = X2 α (n-1) = X2 0.05, 9 = 16.92 Where,O= observed frequencyE= expected frequencyα= level of significance(n-1)= degree of freedomX2= goodness of fit, andH0 and H1 are null hypotheses.c)DiscussionDecision rule says, reject H0 if X2calculated > X2tabulated. With the two levels of significance (i.e. 1% and 5%) and tabulation done so far, since X2calculated is less than X2tabulated hence, we do not reject H0but conclude that Sophisticated IDEs do affect learning strength of newprogramming students. According to the views of the population understudy, some of the adverse effects of sophisticated IDEs on novices among many include, difficulties with its technical knowhow, time wasted in understanding the complexity of the IDEs, a source of hindrance for learning new languages more so when they also require different IDEs, huge computer system resources consumed and multiple diversification of attentions to studying the IDE concerned and the work to execute.Table 2. Tabulation of X2IV.C ONCLUSIONSThis academic work researched the adverse effect of integrated development environments that are built with complex program development tools on programmers who have just ventured into programming. This research has been done by sampling the opinions of some students and lecturers of programming related disciplines who are believed to be directly concerned. The data were sampled using questionnaires and chi square statistical tool was used for analyzing the collected data.At the end of the research, it could be drawn that the integrated development environment whose purpose of development is to aid the programmers in developing effective and quick software applications might do the reverse if the programmers do not go for the right option. New beginning programmers are encouraged to be using non-complex IDEs at their elementary years of programming while complex IDEs can now be introduced at later years when they have experiences of programming. It will be a word of an advice to recommend a better development of other category of IDE that will bridge the gap between sophisticated and non-sophisticated IDEs. This is assumed to provide novices with little technical features and not advanced features as contained in the sophisticated IDEs.A CKNOWLEDGMENTWe wish to acknowledge the effort of our academic mentor who is also our brother, Owoseni, Joshua O., Lecturer in the Department of Applied Geology, Federal University of Technology, Akure, Nigeria, whose ways of academic life challenge us a lot. One will be indebted if the indefatigable effort and support of Mrs C. O. Akanji of Plant Science and Biotechnology, Adekunle Ajasin University, Akungba, Nigeria is not acknowledged. We also wish to thank our Parents, Siblings, Computer Science Department Coordinator (Interlink Polytechnic, Nigeria), NDII and HNDI Computer Science Students of the above named department and to all that might have contributed directly or indirectly towards the success of this paper. Above all, to God be the glory.R EFERENCES[1]I. T. Chan Mow, “The Effectiveness of CognitiveApprenticeship based Learning Environment (CABLE) inTeaching Computer Programming”. Unpublished PHDdissertation, University of South Australia, 2006.[2]I. T. C han Mow, “Analysis of Student ProgrammingErrors in Java Programming Courses,” Journal ofEmerging Trends in Computing and Information Sciences,(2012), Vol 3, No 5, Page 739-749.[3]Source Code Editor,/wiki/Source_code_editor.[4]Build Automation,/wiki/Build_automation.[5]/2010/12/22/advantage-and-disadvantage-of-using-ide/[6] A. Robins, J. Rountree and N. Rountree, “Learning andTeaching Programming: A Review and Discussion,”Computer Science Education, (2003), Vol 13, No 2, Page137-172.[7]K. Muslu, Y. Brun, R. Holmes, M. D. Ernst and D. Notkin,“Speculative Analysis of Integrated DevelopmentEnvironment Recommendation,” OOPSLA’ 12Proceedings of the ACM international conference onobject oriented programming systems languages andapplications, Page 669-682, ACM New York, NY, USA [8] C. Chen, S. Cheng and J. Mei-Chuen Lin, “A Study ofMissconceptions and Missing Conceptions of Novice JavaProgrammers,”.ar/worldcomp2012-mirror/p2012/FEC2866.pdf, retrieved on 1/07/2015.[9]I. Ruchkin and V. Prus, “Single-window integrateddevelopment environment,” /1207-1493.pdf, retrieved on 1/2/2015.[10]M. M. Mhashi and A. M. Alakeel, “Difficulties FacingStudents in Learning Computer Programming Skills atTabuk University,”International conference, 12th,education and educational technology, recent advances inmodern educational technologies, (2013), Page 15-24. [11]D. Teague and P. Roe, “Collaborative Learning-towards as olution for novice programmers,” ACE’ 08 Proceedingsof the tenth conference on Australian computationeducation, Vol 78, Page 147-153.[12]M. Butler and M. Morgan, “Learning challenges faced bynovice programming students studying high level and lowfeedback concepts”, Proceedings of the 24th asciliteConference, (2007), Page 99-107.[13]Turbo Pascal,/wiki/Turbo_Pascal[14]NetBeans, /wiki/NetBeans[15]/visualstudio[16]Microsoft_Visual_Studio,/wiki/Microsoft_Visual_Studio[17]Eclipse, /wiki/Eclipse_(software)[18]JDeveloper,/wiki/Oracle_JDeveloperAuthors ’ ProfilesMr. A. T. Owoseni , is currently a lecturer at department of Computer Science Interlink Polytechnic, Nigeria. His research interests include multi-valued logic, artificial intelligence (AI), information retrieval, programming logic, software engineering, and database management system. He is currently amember of International Association of Engineers; International Association for Computer Science and Information Technology; and few societies of International Association of Engineers. Mr. S. A . Akanji , an environmental statistician, currently lecturing at the department of mathematics and statistics, Interlink Polytechnic, Nigeria,His research interest includes, Modelling Vehicular Emission in Comparative analysis of Statistical Neural Network and Classical Regression.Howtocite this paper: Alaba T. Owoseni, S. A. Akanji,"Survey on Adverse Effect of Sophisticated Integrated Development Environments on Beginning Programmers' Skillfulness", International Journal of Modern Education and Computer Science(IJMECS), Vol.8, No.9, pp.28-34, 2016.DOI: 10.5815/ijmecs.2016.09.04。

I.J.Modern Education and Computer Science, 2013, 4, 43-48Published Online May 2013 in MECS (/)DOI: 10.5815/ijmecs.2013.04.06Understandings of Graduate Students on Natureof ScienceMustafa Serdar KoksalDepartment of Science Education, Inonu University, Campus, Malatya,TurkeyEmail: bioeducator@Canan Tunc SahinDepartment of Elementary Education, Bulent Ecevit University, Kdz Eregli, Zonguldak, TurkeyEmail: cnntnc@Abstract— Knowing about nature of science (NOS) and its products is a basic requirement of all graduate students and researchers due to being both members of society and experts on different scientific disciplines. As the first step, determining NOS understandings of graduate students has importance to go further in developing current situation. Therefore, this study aimed to determine NOS understandings of graduate students from different disciplines. The study included seven graduate students who were enrolled in universities as researchers. As the data collection way, face-to-face interview was utilized. The data of the study was analyzed by assigning the participants to four categories; expert, naive, mixed and not applicable. The results showed that majority of the participants were expert on social and cultural embeddedness of science and role of creativity and imagination in science while majority of the participants were naive on the aspe cts of ―hierarchy between theories and laws‖. Majority of them had mixed understandings on the aspects of existence of only one method in science, subjectivity, tentativeness. Interestingly, all of the participants were naive in terms of definition of science. The results and implications of the study will be discussed.Index Terms—Nature of Science, Graduate Students, Researchers, Science.I.I NTRODUCTIONScience and science-related products have been incorporated into daily life day by day. This changes requirements of an individual to maintain her/his daily life in more comfortable and satisfying way. Scientific literacy as term has been defining the requirements and qualifications of contemporary science-dependent life. Scientific literacy was determined as an important aim of science education in many curriculums and international examination frameworks [1-5]. Scientifically literate individuals are effective and productive members of a society. According to Hurd, these individuals distinguish experts from the uninformed; distinguish theory from dogma, and data from myth and folklore. They also recognize that almost every fact of one’s life has been influenced in one way or another by science/technology and recognize that our global economy is largely influenced by advancements in science and technology. Differently from scientifically illiterate individuals, they view science–social and personal–civic problems as requiring a synthesis of knowledge from different fields including natural and social sciences and also recognize that science–social problems are generally resolved by collaborative rather than individual action [6]. In parallel to need for such individuals, societies have been adapting their curriculums to educate their children in line with requirements of being scientifically literate [1-2]. Scientific literacy has many aspects including content knowledge, nature of science (NOS) and affective characteristics about science. As the most frequently emphasized aspect, nature of science has been studied for a long time by different researchers [7].Nature of science includes different aspects for science education from scientific method to science in society. The aspects of nature of science are described as the followings; ―no universally acce pted one way to do science‖, ―tentative nature of scientific knowledge‖, ―evidence and observation based science‖, ―importance of creativeness and imagination to produce scientific knowledge‖, ―no hierarchy between theory and law‖, ―social and cultural embeddedness‖, ―subjective science‖ and ―science as a way of knowing‖ [8-9]. As a result of epistemological and educational studies, these aspects have been emphasized to be necessary to teach about nature of science [8]. The studies on these aspects were frequently conducted on pre-service teachers or middle, high school students [10-13]. Therefore, studies with graduate level students or researchers are needed to see whole picture of NOS understandings of individuals who took science courses during formal education and conducted scientific research and to make comparisons between educational levels in terms of gaining appropriate NOS understandings. In spite of small number of the studies, there are examples on the studies focusing nature of science understandings of graduate level students or researchers [14-15].Irez studied with 15 Turkish prospective science teacher educators and he found that majority of the prospective science teacher educators had misunderstandings about various NOS aspects [14]. Byapproaching from epistemology perspective, Chang studied on the epistemological positions of graduate students and found that Taiwanese graduate students in the fields out of education had firmly logical positivist epistemological position about scientific knowledge [15]. In contrast, Jehng, Johnson and Anderson found that graduate students had more sophisticated ideas on tentativeness of scientific knowledge [16] and Paulsen and Wells added that age is also another contributor of difference in epistemological understandings [17]. They stated that the more people are getting older, the more they have sophisticated epistemological understandings. Marzooghi, Fouladchang and Shemshiri also found the change in age as an important factor to explain epistemological differences between younger and older university students[18].As speculated by Paulsen and Wells, and Jehng et al.; understandings about nature of science aspects have been varying toward graduate level education [16-17]. Especially, variation might increase due to more focused and narrower specialization experiences. The most important difference between graduate and undergraduate programs is to get opportunities for studying freely on a more specific field of study. Samarapungavan, Westby and Bodner also stated that graduate students have more agency and control over their doctoral studies and might be independently study on empirical anomalies they have met[19].The graduate students have an important place in nature of science studies due to their authentic science experiences and potential to be scientist in future. Having informed understandings about nature of science is the basic requirement for them to conduct epistemologically and theoretically sound and balanced projects. At the same time, they are also in need of being scientifically literate so that they should also be active member of society and make informed decision making on their daily problems. Therefore, determining their NOS understandings have importance for establishing more comprehensive programs and making needed changes in the graduate programs.Based on the conflicting results on NOS understandings of graduate students and importance of having informed NOS understandings for graduate students, this study proposed to determine NOS understandings of graduate students from different departments.The content of this paper is organized in four parts; research question, methodology, results, discussion, implications and suggestions.II. RESEARCH QUESTIONThe main question of this r esearch is ―What are understandings of graduate students who are conducting scientific research studies in different fields of study on nature of science?.‖III. METHODOLOGYIn this study, qualitative research approach was utilized. The study included seven graduate students who were enrolled in universities as researchers. Table 1 is presenting characteristics of them. The participants were selected by taking into account variation in the disciplines in which they studied. Then, they were conveniently determined. As the data collection way, face-to-face interview was utilized by using the questions of VNOS-C (View on Nature of Science Questionnaire-Form C) developed by Lederman, Abd-El-Khalick, Bell and Schwartz [9]. The answers to the questions were recorded by a voice recorder and then were transcribed verbatim. The data of the study was analyzed by assigning the participants to three categories; expert, naive and mixed. Some of the participants did not provide appropriate data to determine profile of the participants on some aspects of NOS; these were categorized as ―not applicable‖. For the categorization, a pre-determined frame provided by Lederman et al. was utilized [9]. The analysis was conducted by considering the approach used by Khishfe and Lederman, Khishfe and Abd-El-Khalick [11, 20]. The questions of the interview had generic nature, so the participants gave answers on more than one aspect of NOS in one question. In determining the profiles, the participants who answered naively on one aspect in all questions were assigned as naive while the participants who answered informedly on one aspect in all questions were categorized as expert. The participants who provided both expert and naive understandings on one aspect in their answers to all questions were assigned as mixed. To increase trustworthiness of the data, two independent examiners have assigned the participants into profile categories. The percent agreement between the examiners has been found as 81%. The discrepancies have been discussed and the final profiles have been assigned to the participants.TABLE IMAIN CHARACTERISTICS OF THE PARTICIPANTSIV. RESULTS AND DISCUSSIONUnder this title, the results will be presented and discussions will be made by considering relevant literature. Results of the categorization of the participants on the NOS aspects are presented in Table 2.TABLE IITHE PROFILES OF THE PARTICIPANTS ON NINE ASPECTS OF NOSAs seen in Table 2, majority of the participants were expert on social and cultural embeddedness of science and role of creativity and imagination in science while all of them were naive in terms of definition of science. At the same time, majority of the participants were also naive on the aspects of ―hierarchy between theories and laws‖. Majority of them had mixed understandings on the aspects of existence of only one method in science, subjectivity, tentativeness. The other aspects were very differently understood by the participants. The quotations from the answers of the participants on the aspects are presented below. The indicators in the46Nature of Science Understandings of Graduate Studentsparenthesis are illustrating participant code, NOS aspect, location and question number of VNOS-C in order. “[There is only way used in science], yes. We can proceed through only questions. Experiments are not requirements for development of all scientific knowledge, in my field of study, experiments have different [way],[Pt4, Only one method in science, Interview, Q3].‖“Laws are proven things while theories are based on only ideas. Theories are tentative while laws are not changes[Pt6, Hierarchy between theories and laws, Interview, Q5].‖―Observations are related to things we can see in natural environments while inference is [based on] estimations we cannot see[Pt1, Observation and inference, Interview, Q6].‖“To reach different results [by using same data] is related to interest of them [sci entist]. …In production of knowledge; cultural effects of society are in case. … The difference of science from religion and philosophy is objective nature of science[Pt3, Subjectivity, Interview, Q1, Q8].‖“I think there is a place for imagination in al l aspects of life. It is a requirement for improvements. …. Absolutely, scientists have been using their creativity and imagination. I think if everything in a field is planned an in an order, it will be boring and developments are stopped [Pt5, Creativity and imagination, Interview,Q10].‖“If new knowledge is added, new perspectives are developed, [theories can change]. Laws are completely accepted. ... Laws are not changeable[Pt7, Tentativeness, Interview, Q4, Q5].‖―Science is based on more concrete evidence and depends on verification while philosophy depends on questioning, but discriminating science from philosophy is very hard. I think both of them are related to each other [Pt2, Empirical basis of science, Interview, Q1].‖“In production of [scie ntific] knowledge, there is an impact of culture. There is a clear example that Vygotsky and Piaget provided different theories as social constructivism and cognitive constructivism. [ by reflecting their cultural their cultural and social values][Pt3, Social and cultural embeddedness, Interview, Q9].‖―Science is a group of knowledge that is based on theories[Pt2, Definition of Science, Interview, Q1].‖In the literature, there are studies supporting the results of this study. Schwartz and Lederman have studied on 24 scientists who are at different scientific disciplines [21]. The authors have applied questionnaire (VNOS-Sci) and interview techniques to collect data. The results of their study has shown that nearly half of the scientists have had expert understanding on tentativeness while other half of the participants have had naive and mixed understandings on the same aspect. Again, they have shown that 66.7% of the participants have also had expert understandings on role of creativity in science. Similarly to the result of this study, majority of the scientists have presented expert views on socio-cultural effects on science while they have also seen naively a hierarchy between theory and law. As another study supporting this study, Samparapungavan, Westby and Bodner have studied on 13 scientists and 22 graduate students by focusing on enacted epistemologies during their projects. The authors have shown that 77% of graduate students and 54 % of scientists have not provided sophisticated and contemporary definition of science [19]. Koksal has also studied on similar group and the author has found as similar to the results of this study that graduate students have misunderstandings on ―hierarchy between theories and laws‖ and ―definition of science‖ aspe cts of NOS [22].The results of this study have shown that graduate students in different scientific disciplines have presented both similar and different understandings on NOS. In spite of similarities, there is no common understanding among the participants; this can be explained by both their difference in experience on scientific studies which are specific to their fields and personal experiences on science. Personal experiences are very effective factors in shaping understandings of individuals on science. Similarity in their understandings on definition of science might be explained by common science curriculum for all children in Turkey and stereotyped definitions of science in Turkish science textbooks [23]. As another similarity in understandings, hierarchy between theory and law might be related to previous exposure to naïve definitions of these terms in different resources. Taskın et al. have studied on theory concept with the 572 undergraduate students and they have found the participants have believed existence of a hierarchy between theory and laws. Then, the authors have asked the participants; ―what can be resources for your naïve understandings‖, the participants have answered the question by referring to false teacher definitions and naïve definitions in textbooks [24].V.DISCUSSION, IMPLICATIONS ANDSUGGESTIONSThe results of this study have showed that varied graduate experiences in different fields are not enough to improve NOS understandings of graduate students. The participants have expert understandings on only two aspects; creativity and imagination in science and effect of socio-cultural factors on science. Majority of the answers have been coded as mixed. This is an indication of lack of improved understandings on the NOS aspects to use in informed decision making. The group of the study needs to be expert in understanding on NOSaspects due to their continuous involvement in scientific process and being informed decision maker on daily life situations. Results of this study refer to need for elimination of naive understandings on NOS at graduate level. The explicit-reflective implications embedded in graduate courses might be helpful to increase expert understandings of the participants [10-11]. As another suggestion, enacted epistemologies of the participants might be shown by providing continuous feed-back on the aspects during a project conducted by graduate students [19].When look at the individuals, it has also been seen that experience in research and taking any course on and participating in any activity related to epistemology, history and nature of science are not effective factors to improve NOS understandings. Moreover, the participant who is graduate student in philosophy department has also presented naive views similar to other graduate students.As an interesting point, none of the participants as active members of scientific research community have defined science in a manner that contemporary definition accepts it. As a suggestion, the effectiveness of courses or activities on NOS teaching should be improved by using individual means such as inserting a NOS reflection part into personal lab report format. At the same time, the graduate courses should include contemporary critics on science and its aspects. Variation in understandings has been showing that the participants have presented both similarities and tendencies to understand some aspects and discrepancies on a different group of the aspects. The most varied profiles have been provided for observation and inference difference and empirical basis of science. This result might be related to difference in observation ways of different scientific disciplines. For example, zoologists use direct observation while a molecular biologist uses indirect ways of observation; reaction time differences. Similarly, frequency of need to make inferences might be different in two different disciplines. Frequently a zoologist can directly classify animals while molecular biologists need to make inferences to classify an animal by using molecular evidence. For the other aspects, the participants have presented less varied understandings. This might be an evidence for independence of NOS understandings from scientific discipline. Only one of the participants (Pt5) has expert understandings on over half of the aspects. It might be related to individual experiences and context in which she has encountered.Based on the results of this study, it can be suggested that the study should be extended by using more comprehensive analysis techniques such as phenomenography to establish individual cognitive maps of the individuals based on their personal experiences. As another suggestion, the future studies should measure epistemological beliefs of the participants and think of NOS and epistemological beliefs together to see roots of the relationship between individual experiences and NOS understandings. The study has a limited number of the participants; therefore, there is a need to reach more participants in each discipline of science. By this way, discrepancies in individual contexts for the same discipline can be explained.This study has some limitations, so the interpretation of the results should be made carefully. The participant number of the study is limited to seven. At the same time, the data resources are limited to interview answers on the aspects of NOS. As another limitation, the aspects which are studied are also limited to nine aspects of NOS. The future studies should take these points into account.REFERENCES[1]Turkish Ministry of Education. Ninth GradeBiology Curriculum. Ankara, Turkey.2007.[2]AAAS. Benchmarks for Science Literacy, Project206, Science Literacy for a Changing Future. Report of American Association for the Advancement of Science. Retrieved on November 2012 from (/). 1994.[3]Kjaernsli, M. and Lie, S. PISA and scientificliteracy: Similarities and differences between the Nordic countries. Scandinavian Journal of Educational Research, 48(3), 271–286.2004.[4]Olsen, R.V., Kjærnsli, M. and Lie, S. Using singleitems in PISA to explore international diversity in scientific literacy. Paper presented at the 5th ESERA conference Barcelona, Spain. 2005.[5]OECD. The PISA 2003 assessment framework:Mathematics, reading, science and problem solving knowledge and skills. Paris: OECD. 2003.[6]Hurd, P.D. Scientific Literacy: New Minds for aChanging World. Science Education, 82 (3),407-416. 1998[7]Lederman, N. G. Nature of science: Past, present,and future. In Abell, S. & Lederman, N. (Eds.) Handbook of Research on Science Education.Mahwah, New Jersey: Lawrence Erlbaum Associates, Publishers. 2007.[8]McComas, W. F. The Principle Elements of theNature of Science: Dispelling the Myths. In W.F.McComas (Ed.), The nature of science in science education: Rationales and strategies (pp. 53–70).Dordrecht, the Netherlands: Kluwer Academic Publishers. 1998.[9]Lederman, N.G., Abd-El-Khalick, F., Bell, R.L. andSchwartz, R. S. Views of Nature of Science Questionnaire: Toward Valid and Meaningful Assessment of Learners’ Conceptions of Nature of Science. Journal of Research in Science Teaching,39 (6), 497-521.2002.[10]Khishfe, R and Lederman, N. Teaching Nature ofScience within a Controversial Topic: Integrated versus Nonintegrated. Journal of Research in Scıence Teaching, 43, 4, 395–418. 2006.[11]Khishfe, R., and Abd-El-Khalick, F. Influence ofexplicit and eeflective views versus implicit inquiry orientated instruction on sixth graders views of thenature of science. Journal of Research in Science Teaching, 39(7), 551-578.2002.[12] Ryan, A. G. and Aikenhead, G.S. Students’Preconceptions about the Epistemology of Science. Science Education, 76 (6), 559-580.1992.[13] Palmquist, B. and Finley, F. Preservice Teachers’Views of the Nature of Science During A Postbaccalaureate Science Teaching Program. Journal of Research in Science Teaching, 34, 595-615.1997.[14] Irez, S. Are we prepared?: An Assessment ofPreservice Science Teacher Educators’ Beliefs About Nature of Science. Science Teacher Education, 90, (6), 1113-1143.2006.[15] Chang, T. An investigation of Taiwanese graduatestudents’ beliefs about scientific knowledge. Bulletin of National Taiwan Normal University. 40, 583–618, 1995.[16] Jehng,J.J. Johnson, S.D. and Anderson, R.C.Schooling and students’ e pistemological beliefs about learning, Contemporary Educational Psychology, 18, 23–35. 1993.[17] Paulsen, M. B. and Wells, C. T. Domain differencesin the epistemological beliefs of college students. Research in Higher Education, 39,(4), 365–384.1998.[18] Marzooghi, R., Fouladchang, M. and Shemshiri, B.Gender and Graduate level differences in epistemological beliefs of Iranian undergraduate students. Journal of Applied Scences. 8, (24), 4698-4701. 2008.[19] Samarapungavan, A., Westby, E. L. and Bodner, G.M. Contextual Epistemic Development in Science: A Comparison of Chemistry Students and Research Chemists. Science Education, 90 (3), 468-495.2006. [20] Khishfe, R and Lederman, N.G. Relationshipbetween instructional context and views of nature of science. International Journal of Science Education, 29 (8), 939-962.2007.[21] Schwartz, R. S. and Lederman, N.G. What scientistssay: Scientists’ views of nature of science and relation to science context. International Journal of Science Education, 30(6), 721-771.2008.[22] Koksal, M.S. Discipline dependent understandingsof graduate students in biology education department about the aspects of nature of science, Education and Science, 35, 157, 68-83. 2010.[23] Irez, S. Nature of science as depicted in Turkishbiology textbooks. Science Education, 93(3), 422-447.2008.[24] Taskın, O. Çobanoğlu, E. O. Apaydın, Z.Çobanoğlu, İ.H.Yılmaz, B. and Şahin, B. Undergraduate Students’ Perception of Theor y Concept. Boğaziçi University Journal of Education, 25 (2), 35-51.2010.Dr. Mustafa Serdar Koksal is an assistant professor at the Department of Science Education, Inonu University. In teaching, he has been focusing on using explicit-reflective approach for teaching NOS. In research, his current interests include epistemological beliefs, risk taking and argumentation. Dr. Kong received his PhD degree from Faculty of Education in Middle East Technical University. He is a member of ESERA and IRATDE.Canan Tunc Sahin has degrees of BE and MS from Bulent Ecevit University Turkey. She has been working for faculty of education over 5 years as a research assistant. She is currently a PhDstudent in Marmara University, Istanbul, Turkey.。

I.J.Modern Education and Computer Science, 2009, 1, 19-27Published Online October 2009 in MECS (/)Analyzing the Influencing Factors of Group Learning: A Mixed ApproachJianhua ZhaoSchool of Information Technology in Education, South China Normal University, Guangzhou, 510631, ChinaEmail: jhuazhao@Yinjian JiangSchool of Foreign Languages, Guangdong Polytechnic Normal University, Guangzhou, 510665, ChinaEmail: georgina29@Abstract—The purpose of this paper is to explore which factors influence group learning content, and content analysis is chosen as the research method. The sample for this study is the literature of group learning. 35 books and 1 paper was examined. The coding system for the content analysis is an opened and a self-expanded system in this study, which means that the original coding system can be updated if the new coding item is developed during the data collection. A total of 62 influencing factors are identified in terms of the content analysis. In order to organise them systematically, we categorised them into four aggregations according to one model of the group learning processes: planning, organising, learning process, and evaluation. The result of this study may be used to design a questionnaire and to model group learning process in our further research.Index Terms—group learning, influencing factors, coding system, content analysisI.I NTRODUCTIONMuch research has already been carried out into group learning, such as group problem solving [1][2], computer-supported group learning [3][4], cooperative learning group [5][6][7], and virtual group learning [8]. When one reviews the work of the field, one can easily find that there has only been a few studies which address the factors that influence group learning [9][10]. To know which factors will influence group learning, it is important for the field of e-learning research, especially for this study, which will focus on how to use computer to facilitate group learning.The factors influencing the performance and effectiveness of a group learning process were defined as the influencing factors of group learning in this study, such as group task, group composition, group communication, group interaction, group structure, and group evaluation. McGrath proposes a paradigm to analyze group interaction (see Figure 1) [11].“INPUT → PROCESS → OUTPUT” for analyzing the roles of the group interaction process. The INPUT component includes individual-level factors, group-level factors, and environment-level factors. The OUTPUT component includes performance outcomes and other outcomes [11]. This paradigm manifests the different influencing factors of group learning. However, it does not clarify how many influencing factors are involved in the different components. Hackman and Morris develop a framework to explain the relationship among the focal input variables, the group interaction process, the summary variables, the critical task contingencies, and group performance and effectiveness (see Figure 2) [12]. Three categories of variables are involved in this paradigm, i.e. effort, performance strategies, and knowledge and skill. Hackman and Morris consider these three categories of variables are the most proximal causes of group task effectiveness. Similar with McGrath’s paradigm, this framework can be described as an “Input-Process-Output” sequence for different types of tasks [11].These variables can be considered as the influencing factors of group learning process in this study, i.e. group composition, group norms, group task, group performance, and group interaction. However, they do not explicate how many influencing factors are involved in this process. Other researchers, such as Jaques and Reynolds, also mention the influencing factors of a group learning process [9][10]. However, they do not introduce the influencing factors of them systematically.In this study, we are going to explore which factors influence group learning processes. In order to identify these factors, we compare some related field works [13][14][15] and choose content analysis as our research method. Content analysis is a research technique for making replicable and valid inferences from data to their context [14]. Conventionally, content analysis can be considered as a qualitative method. But Berg argues that “content analysis can be considered as a blend of qualitative and quantitative method” [15]. We prefer this perspective. In this study, we will use a qualitative method to identify the factors first, and then use quantitative methods to extract which factors are more essentially related to group learning processes. By thiswe suggest that these factors will influence group learning process more.Figure 1. A traditional paradigm for analysis of group interaction as a mediator of performance outcomes Figure 2. An input-process-output sequence framework for different types of tasksIt is important to know explicitly the relevant influencing factors of a group learning process in order to analyze and utilize them in-depth. These factors are related to the different stages of group learning process, such as group organizing, group process, and group performance. If the influencing factors of a group learning process were identified, they could be used to devise and manage this process intentionally. If a tutor was familiar with these influencing factors, it could help her/him to organize a group learning process effectively for her/is teaching.Two approaches could be employed to find out the influencing factors of group learning process. The first approach was to analyze the actual group learning processes and to extract the relevant influencing factors via a series of experiments. However, it was a time-consuming approach and difficult to ensure these influencing factors could be extracted completely. The second approach was to analyze them from relevant literature. The relevant research had already been undertaken in the field. To analyze the literature would help us to effectively find out these influencing factors.II.M ETHODOLOGYINPUT PROCESS OUTPUTTIMEt12A. Content AnalysisContent analysis is a research technique for making replicable and valid inferences from data to their context [16]. Its purpose is to provide knowledge, new insights, a representation of “facts”, and a practical guide to action. McKernan states that content analysis focuses on the inquiring into the deeper meaning and structure of a message or communication [17]. The messages may be contained in a written document, a communications broadcast, film, video, or in actual human behavior observed. The purposes are to explore the hidden themes, concepts, and indicators of the message contents. Therefore, content analysis can be considered as a tool that can be used to explore the specific data. Researchers will examine artifacts of social communication in content analysis, such as documents, transcriptions, and videotapes.Normally, the data of content analysis is qualitative, rather than quantitative. Therefore, it is a qualitative method. The qualitative aspect of content analysis includes examining ideological mind-sets, themes, topics, symbols, and similar phenomena. However, Berg argues that content analysis can be considered as a blend of qualitative and quantitative method [18]. A series of tally sheets to determine specific frequencies of relevant categories will link with its quantitative aspect. This study views content analysis as a blended method, which is quite similar to Berg’s viewpoint.The purpose of content analysis in this study could be summarized as: to analyze the relevant literature and to find out the influencing factors of a group learning process. These factors could be used to design the specific questionnaires to examine the differences of group learning in the classroom-based and the web-based settings.The data for content analysis in this study includes two categories, i.e. quantitative and qualitative data. The qualitative data was collected from the relevant literature, i.e. books, and journal papers about group learning. These materials provided the systematic and synthetically perspectives on analyzing group learning process.A coding system was developed in order to use content analysis first. The primary coding system was developed according to analysis of McGrath’s model and Hackman and Morris’s model [11][12]. This coding system was an open system, which means that the new influencing factors of a group learning process could be identified and added to this coding system during the content analysis process.Data analysis provided the frequency and percentage of the influencing factors in the literature, which were the quantitative data of content analysis. The influencing factors were extracted according to the significance of their frequency and percentage.B. SamplingThe research about group learning in the literature presented its social, psychological, political, and educational characteristics, which formed a group learning research community in the field. The outcomes and achievements of this community could be considered as the population of this study, such as books, papers (journals and conferences), reports, or theses. The relevant issues of this study included group performance, group interaction, group conflict, group leadership, group role, group communication, group dynamic, group structure, group process, and group work. These issues represented the population of this study.Group research in literature is already related to its various aspects, such as social, psychological, political, educational characteristics. The samples in this study are chosen from typical books and journal papers in the group learning research field. The issues include group performance, group interaction, group conflict, group leadership, group communication, group dynamic, group structure, group role, group process, and group work, which represent the essential of group. The influencing factors of group learning can be extracted from these issues by content analysis. The name of each sample can be defined according to their original book title. For this purpose, we define a set of names for the samples which are described in Table I.TABLE I.T HE NAME OF THE SAMPLESNo The Name ofSamplesThe Title of the LiteratureThe type ofSamples1CL Communication and Learning in Small Group B2LG Learning in Groups B 3WG Working in Groups: Communication Principles and Strategies B 4SG Small Group Learning in the Classroom B 5GP Group Process: Papers from Advances in Experimental Social Psychology B 6GT Group Theory for Social Works: An Introduction B 7GA Groups at Work B 8 GR Groupwork B 9GK GroupworkPractice B 10DG Dynamics of Group Action B 11MG Motives and Goals in Groups B 12LE Learning from Others in Groups: Experimental Learning Approaches B 13GD Group Dynamics & Individual Development B 14IC Interaction in Cooperative Groups: The Theoretical Anatomy of Group Learning B 15GTC Group Tutoring, Concepts and Case Studies B16JT Join Together: Group Theory & Group Skills B 17IB Intergroup Cognition & Intergroup Behavior B 18DM Group Decision Making B 19AGP Socio-Psychological Aspect of Group Process J 20 GC GroupDynamics B 21LP Group Work: Learning and Practice B 22LTS Learning Through Small Group Discussion: A Study of Seminar Work in Higher Education B 23DWB Group Process: Dynamics Within and Between Groups B 24SGW Successful Group Work B 25CMC Cooperation in the Multi-Ethnic Classroom B 26 GPE GroupPerformance B 27CST Communication in the Small Group: Theory and Practice B 28GPC Group Process in the Classroom B 29GET Groupwork in Education and Training: Ideas in Practice B 30GDR Group Dynamics: Research and Theory B 31ISG Interaction in Small Groups B 32HSG Handbook of Small Group Research B 33LTG Learning Through Group Experience B 34TGT T-Group Theory and Laboratory Method: Innovation in Re-education B 35MTM Modern Theory and Method in Group Training B 36STD The Structure and Dynamics of Organizations and Groups BIn the Table I, we list the sequence number, the name, the title, and the types of samples. From the sequence number, we can know that the total number of the samples in this study is 36. In ‘The Name of Samples’ column, we define the different names for the samples. ‘The title of literature’ is the title of each book or academic published papers in journals. There are two types’ samples in our content analysis, which are listed in the column of ‘The types of samples’. ‘B’ means book and ‘J’ means journal. From table I, we can know there is only one sample was chosen from the journal.We organize these samples into different categories in terms of their main characteristic. The categories and the distribution number are listed in Table II.TABLE II.C ATEGORIES OF THE SAMPLESNo. Categories The Name of the Samples DistributionNumberPercentage (%)1Group Communication CL, WG, CST, 3 8.332 GroupPerformance GPE 1 2.783Group Interaction IC, ISG 2 5.564Group Process GP, AGP, DWB, GPC 4 11.115Group Learning LG, SG, LE 3 8.336Group Dynamics DG, GD, GC, GDR, STD 5 13.897Group Work GT, GA, GR, GK, LP, SGW, CMC, GET 8 22.228 T-Group TGT,MTM 2 5.56 9 GroupExperience LTG 1 2.7810 GroupMotives MG 1 2.7811 GroupTutoring GTC 1 2.7812 GroupSkills JT 1 2.78 13 Inter-groupBehaviour IB 1 2.7814Group Decision Making DM 1 2.7815 GroupDiscussion LTS 1 2.7816 Others HSG 1 2.78The samples can be categorized into 16 issues according to their main topic, which are listed in the column of ‘the name of the samples’ in Table II. From ‘distributed number’, we know how many samples are included in each category. ‘Percentage’ gives us the further information about the ratio of each category in the total samples. The maximum category of sample is ‘group work’ (22.22%). ‘Group dynamic’ (13.89%) is the second maximum category. ‘Group process’ (11.11%) is the third maximum. ‘Group communication’ (8.33%) and ‘group learning’ (8.33%) are the fourth maximum. ‘Group interaction’ (5.56%) and ‘T-group’ (5.56%) are the fifth maximum. Others (2.78%), which include ‘group performance’, ‘group experience’, ‘group motives’, ‘group tutoring’, ‘group skills’, ‘inter-groupbehavior’, ‘group decision making’, ‘group discussion’, and ‘others’ are the sixth.The publishing dates of these samples were presentedin Table III. The time span was from 1962 to 2000 (38 years), which represented a long term perspective on group learning. Field research was most active in 1978, 1984, and 1994. Four samples had been published respectively in each of these three years (11.11%).The 1970’s was the most active decade with twelve samples were published, followed by the 1990’s with nine samples. The 1960’s and 1980’s were quite similar to six or seven books each.TABLE III.THE TIME DISTRIBUTION OF THE SAMPLESYear Samples Dist.NumberDist.NumberPercentage(%)1962 HSG6 1 2.781963 STD 1 2.78 1964 DG,TGT 2 5.56 1966 LTG 1 2.78 1968 GDR 1 2.781971 MG12 1 2.781972 MTM 1 2.78 1974 GD 1 2.78 1975 ISG 1 2.78 1976 GK 1 2.781978 LTS, GPC,LP, GP4 11.111979 GTC, LE,GR3 8.331981 GA7 1 2.781984 DM, GT,LG, CL4 11.111989 CST,SG, 2 5.561990 GC9 1 2.781994 CMC, GET,GPE, JT4 11.111995 AGP,IC 2 5.56 1996 SGW 1 2.78 1998 IB 1 2.78The primary coding system is developed according to the relevant research on the influencing factors of group learning. The fundamental factors are involved in this primary coding system. The coding system includes three main contents, i.e., code, definition, and example. This primary coding system is an open-ended system which means the extra influencing factors can be addedto the coding system when is identified from the literature. The developing process of coding system includes three steps, i.e., developing the primary coding system, new influencing factor is added into the primary coding system, and developing a relative coding systemFigure 3. The process to construct the norms of coding systemStep 1: Constructing the primary coding systemThe primary coding system is built to analyze a common group learning process, which is the foundationto build a coding system. Certainly, there are some influencing factors which are connected with group learning process, which are easily gathered through simply analysis. We deal with this work in terms of a framework described by Hackman and Morris [12]. They considering that there are three classes of variables, e.g., effort, performance strategies, and knowledge and skill, are the most powerful proximal causes of group task effectiveness. These variables can be expressed in an ‘input-process-output’ sequence for different types of tasks (see Figure 2).The primary coding system can be built according to an analysis this framework, which includes some influencing factors to the group performance effectiveness, such as ‘group composition’, ‘group norms’, ‘group task’, ‘group interaction’, ‘group strategies’, ‘group performance’, ‘group effectiveness’, ‘group outcomes’, and ‘group design’. However, the ideal coding system cannot be acquired according to this way. The primary coding system need following remedy. Step 2: Adding new factors into the primary coding systemWe define our coding system as a self-expanding and self-maturing system, which means that it can be mended according to the addition of new factors into the primary coding system. These factors can be analyzed and collected from the specific samples (books and journals). It means that this coding system is opened and not a ready-made coding system can be used into this work. Therefore, this process also can be considered as the way to build a coding system for content analysis.Step 3: Refining and elaborating the coding systemThe primary coding system can be refined and elaborated after continual remediation. The relatively complete coding system will be built during our content analysis. Definitely, if we want to improve the precision of the coding system, further remediation also is needed.III.D ATA C OLLECTIONThe data in this research was gathered from the samples through content analysis in terms of the coding system. In order to collect data, we designed a form and named it “Data Collection Form via Content Analysis,” which was devised in terms of the coding system. Each books or paper uses one form for its data collection. The new influencing factors are added in this form and also are added into the coding system.Generally, data collection in content analysis should be guided by a coding system. Moreover, participants need to be trained first in order to in-depth fully comprehend the coding system. In this study, because there is no predefined coding system, we do not need to collect data according to the general way. Our work in data collection includes three steps.Step 1: Quoting the relevant sentences or paragraphsfrom the literatureReading the books and journals (the chosen samples) in order to quote the paragraphs when it expresses the meaning which is connected to the effectiveness of group learning process. When they are identified, they are then quoted. The original materials from quotation work includes the index number (CA00**, CAB**, CAC**, and CAD**), quoted content, page number, andinfluencing factors. This work can be done partly depend on the primary coding system. The paragraph will be quoted when it meet the norm of coding. For instance: CAB008: We suggest that the key to understanding the ‘group effectiveness problem’ is to be found in the on-going interaction process which takes place among group members while they are working on a task. P2 ---- [19]CAC028: As the end of the period of observation it is possible to collate the observations in each category and provide an interaction profile of the group as a whole (in terms of the percentages of time spent engaged on the different categories of behavior), of individuals in the group, or (the most complete picture of all) the proportion of time each person spent interacting with the others and in what manner. P42 ---- [20]When the quotation work is finished, these original quoted materials can be used to analyze the relevant influencing factors according to the coding system.Step 2: Analyzing the sentences or paragraphs and to identify or extract the influencing factors of group learningTick in the data collection form in terms of the quoted paragraph from the samples. When the quotation work is finished, the original quoted materials can be used to tick mention highlight factors in the data collection form. The frequency of factors does not need to be calculated in the same book or journal. Therefore, the mentioned factors in one book or journal only are recorded one times. Some examples demonstrate how we recognize the influencing factors for this purpose, for examplesCA0006: A supplement to Smith’s paper is my own chapter summarizing some group behavior theories for community workers. P14 ------ (Group behavior) [21]The author introduces how he summarizes group behavior theories in his own chapter. I recognize it is related to group behavior. Therefore, I quote it out and categories it as group behavior.CAB029: In summary, these studies suggest that the impact of group interaction on group performance can be analyzed systematically and that the results of such analyses can increase understanding of the reasons why some groups are more effective than others. P8 --- (group interaction, group performance) [19]This paragraph introduces the results of authors’ studies. From this paragraph, I can extract group interaction and group performance.CA0032: Group life explored in this way may be a liberating experience but often proves to be a dislocating one for individuals when they have torelate to people who have not been expose to the same liberating experience, and have to discover in that context how to be more authentically themselves and transform habitual role behaviors. P31 ---- (Group life, group experience, group setting, role behaviors) [21] From this paragraph, I can identify some influencing factors of group learning, i.e., group life, group experience, group context, and role behaviors.CA0114: It has been found that groups comprising between ten and thirteen participants, including one consultant, provide optimum conditions for learning about group processes in a simple social setting. P87 ---- (group composition, group size, group context, group processes, group community, and role playing) [21]Researcher introduces group composition which includes group size, role playing, and group context in a simple social setting (community). From this paragraph, I can extract few influencing factors of group learning, i.e., group composition, group size, group context, group processes, role playing, and group community.CA0211: Developing the self-confidence of members in their ability to fact authority figures and to take some control of their own environment can be helped in the group through the planning of occasions where the group negotiates about rules or resource or invites visitors with resources into meetings. P186 - 187 ---- (Individual ability, Group environment, Group Negotiation, group rules, Group Resources, group meeting) [21]This paragraph introduces how to develop the self-confidence of members. The influencing factors of group learning can be identified as individual ability, group environment, group negotiation, group rules, group resources, and group meeting.CAB030: One hundred and eight experimental groups spent 15 minutes on each of four ‘intellective’ tasks. Four hundred and thirty-two separate transcripts of group interaction and 432 group products were obtained. A total of 144 different group tasks were used in the research, 48 each of three task ‘types’: (1) ‘production’ tasks, which require the production and presentation of ideas or images; (b) ‘discussion’ tasks, which require and evaluation of issues; and (c) ‘problem-solving’ tasks, which require specification of a course of action to be followed to resolve some problem. P9 ---- (group interaction, group tasks, group discussion, group problem-solving, group evaluation) [19]This is a long paragraph. The author summarizes the findings of their study on group learning. Lots of influencing factors of group learning can be found out from it, i.e., group task, group interaction, groupproduction, group discussion, group problem- solving, and group evaluation.CAB194: At the end of this first set of group trials, group performance was scored and points based on the group leader. P127 --- (group performance, group leader) [19]From this sentence I can identify two influencing factors of group learning, i.e., group performance and group leader.When the influencing factors are identified, write them down on the each sentence or paragraph. Then, tick in the data collection form in terms of the results of the identification. I do not analyze the frequency and percentage of influencing factors in the same book or journal. Therefore, a certain influencing factor is found in one book or journal paper, it was just ticked one time. Step 3: Adding new influencing factors into the data collection formWhen a ‘ticked’ work meets the situation where the factor cannot be found in the data collection form, in this case, the factor will be added into the data collection form. The blank cells in the form can be used for this purpose. Meanwhile, the new factor also needs to be added in the coding system, and furthermore, it can be considered as the rule which will be used to deal with the other samples.IV. D ATA A NALYSISIn order to approach the further analysis, the data collected through content analysis should be given a brief introduction first. We use SPSS as a processing tool to get the data summary. The frequency of each factor can be described in Figure 4.510152025303540G B HG I NG C OG E TG S TG D EG A CG S T RG R EI N RG A TG C O NG R OG M IG C AG G OI N CG C YG D I SG P SG P SG H IG C RG D IG E XI A TG N EG M OR P LG F OG M TI B EG E YG A NG A BG M UG S PG P NFigure 4. The frequency of influencing factorsIn Figure 4, the range of frequency is 35. 13 factors are less than 10, which are ‘group duration’ (GDU, 5), ‘individual goal’ (IGO, 8), ‘individual decision’ (IDE, 1), ‘individual behaviour’ (IBE, 9), ‘group efficiency’ (GEY, 6), ‘individual performance’ (IPE, 1), ‘group ability’ (GAB, 6), ‘group consciousness’ (GCS, 7), ‘group movement’ (GMV, 5), ‘group controversy’ (GCT, 4), ‘group progress’ (GSP, 6), ‘group think’ (GTH, 8), and ‘group presentation’ (GPN, 6). 12 factors are equal to 36, which mean they got full recognized. These factors are GBH (group behaviour), GCO (group communication), GTA (group tasks), GDE (group decision), GSTR (group structure), GSI (group size), GEN (group environment), GCON (group conflict), GGO (group goals), GRES (group resources), GDIS (group discussion), and GLE (group leader). 12 factors are more than 10 (included)and less than 20. 13 factors located 20 (included) and 30. Other 25 factors are more than 30 (included) and less than 36.The number of frequencies is less than 10 does not mean it cannot be used and less validity. It just means that it was referred not too much in these chosen samples. The number is bigger, which means it got more concern by field researchers. In this study, we also need pay more concern and well-analysis for these factors.V. R ESULTS AND C ONCLUSIONWe can get the influencing factors of group learning process from the analysis on literature, and present them in Table IV. There are 62 influencing factors in total.TABLE IV.T HE INFLUENCING FACTORS OF GROUP LEARNING PROCESSNo. Name Freq. Freq.1 GFE (Group feedback) 3332GCN (Group contribution) 102 GBH (Group behaviour) 36 33INR (Interpersonal relationships) 323 GCO (Group communication) 3634GAT (Group attitudes) 324 GDE (Group decision) 3635GSK (Group skills) 325 GSTR (Group structure) 3636GDY (Group dynamics) 326 GSI (Group size) 3637GRE (Group rewards) 307 GEN (Group environment) 3638GME (Group methods) 308 GCON (Group conflict) 3639RIN (Relationship of inter-group) 299 GGO (Group goals)3640GST (Group strategies) 28。

神经网络及深度学习(包含matlab代码).pdf
神经网络可以使用中间层构建出多层抽象，正如在布尔电路中所做的那样。

如果进行视觉模式识别，那么第1 层的神经元可能学会识别边；第2 层的神经元可以在此基础上学会识别更加复杂的形状，例如三角形或矩形；第3 层将能够识别更加复杂的形状，以此类推。

有了这些多层抽象，深度神经网络似乎可以学习解决复杂的模式识别问题。

正如电路示例所体现的那样，理论研究表明深度神经网络本质上比浅层神经网络更强大。

《深入浅出神经网络与深度学习》PDF+代码分析
《深入浅出神经网络与深度学习》PDF中文，249页；PDF英文，292页；配套代码。

提取码: 6sgh
以技术原理为导向，辅以MNIST 手写数字识别项目示例，介绍神经网络架构、反向传播算法、过拟合解决方案、卷积神经网络等内容，以及如何利用这些知识改进深度学习项目。

学完后，将能够通过编写Python 代码来解决复杂的模式识别问题。

使用Deeplearning4j进行语义分割语义分割是计算机视觉领域的一个重要任务，旨在将图像中的每个像素分配给特定的语义类别。

Deeplearning4j是一个基于深度学习的开源库，可以用于实现语义分割任务。

本文将详细介绍如何使用Deeplearning4j进行语义分割。

1. 简介语义分割是图像分割的一种特殊形式，旨在实现对图像中每个像素的语义内容进行标记。

与传统的图像分类任务只需要为整个图像分配一个类别不同，语义分割需要为每个像素分配一个类别标签，从而实现对图像细粒度的理解。

2. Deeplearning4j的安装与配置首先，我们需要安装Deeplearning4j库。

可以通过pip命令来安装Deeplearning4j:```pip install deeplearning4j```安装完成后，我们可以通过import语句导入Deeplearning4j库并开始使用。

3. 数据准备在进行语义分割之前，我们需要准备标注好的训练数据。

训练数据包括原始图像和对应的像素级标注，标注为每个像素指定相应的类别。

可以使用图像标注工具来手动标注数据，也可以使用现有的数据集。

4. 构建模型Deeplearning4j提供了一系列用于语义分割的预训练模型，如U-Net、SegNet等。

在构建模型之前，我们需要定义模型的架构、超参数等。

```pythonfrom dl4j.models.unet import UNet# 定义模型架构model = UNet(num_classes=10)# 设置超参数pile(optimizer='adam', loss='categorical_crossentropy')```5. 训练模型在准备好数据和定义好模型之后，我们可以开始训练模型。

通过调用模型的fit方法，可以将准备好的训练数据作为输入进行训练。

```pythonmodel.fit(X_train, y_train, batch_size=16, epochs=10,validation_data=(X_val, y_val))```6. 预测结果训练完成后，我们可以使用训练好的模型对新的图像进行语义分割。

incremental learning 代码
增量学习（Incremental Learning）是一种机器学习技术，它允许模型在新的数据到达时逐步学习并改进。

这种方法通常用于大规模、高维度的数据集，如互联网图像、社交媒体数据等。

以下是一个简单的增量学习代码示例，使用Python和Scikit-learn库：
python复制代码
from sklearn.ensemble import RandomForestClassifier
import numpy as np
# 初始化模型
model = RandomForestClassifier()
# 初始化数据集
X = np.array([[1, 2], [3, 4], [5, 6]])
y = np.array([0, 1, 0])
# 训练模型
model.fit(X, y)
# 添加新数据
X_new = np.array([[7, 8], [9, 10]])
y_new = np.array([1, 0])
# 训练模型
model.partial_fit(X_new, y_new)
在这个示例中，我们首先使用随机森林分类器初始化模型。

然后，我们使用初始数据集训练模型。

接下来，我们添加新的数据并使用partial_fit方法训练模型。

partial_fit 方法允许我们在不重新训练整个模型的情况下添加新的数据，因此它可以更快地适应新数据。

这只是一个简单的示例，实际增量学习的实现可能更加复杂。

具体实现取决于使用的算法和数据集。