Combining Classification and Model Trees for Handling Ordinal Problems

代数中的英语词汇1、Numbers-数Natural Numbers—自然数Zero—零Integers—整数Negative integers—负整数Decimal—小数Fractions—分数['ræʃənəl]Rational Numbers—有理数[i'ræʃənəl] Irrational Numbers—无理数Real Numbers—实数[i'mædʒinəri]Pure Imaginary Numbers—纯虚数['kɔmpleks] Complex Numbers—复数Sets of Numbers—数集2、Integers—整数['intidʒə]Expression of Integers—整数的表达Grouping of Integers—整数编组Units or Ones—个Tens—十Hundreds—百Thousands—千Ten Thousands—万Hundred Thousands—十万['miljən] Millions—百万Ten Millions—千万Hundred Millions—亿Billions—十亿Ten Billions—百亿Hundred Billions—千亿Odd Numbers—奇数Even Numbers—偶数Negative IntegersOrigin—原点Positive direction—正向Negative direction—负方向Minus—负的Powers of Integers—整数的幂['kɑ:dinl]Cardinal Numbers—基数['ɔ:dinəl]Ordinal Numbers—序数['ærəbik]Arabic English—阿拉伯英语Rounding-off Integers—四舍五入3、Decimals and Fractions['desiməl] Decimals—小数['frækʃən]Fractions—分数Mixed Decimals—混合小数[ə'prɔksimit] Approximate Numbers—近似数[sig'nifikənt] Significant Digits—有效数字[nəu'teiʃən] Scientific Notation 科学计数法Power Expression of Decimals小数的幂形式Quarter—四分之一Conversion of Fractions and Decimals--分数和小数之间的变换[pə'sent]Percent—百分数[pə'mil]Permille—千分数4、Signs-符号Basic Mathematical Signs—基本数学符号=equals—等于Signs of Operation—操作符+ Plus; Positive—加，正的— Minus; Negative—减，负的Multiply乘multiplied by; times—乘over; divided by—除[prə'pɔ:ʃən]Proportion —比例['reiʃiəu]Ratio—比率Percent—百分比[brækəts][ ]Brackets—括号，方括号( ) Round Brackets—圆括号[pə'renθisi:z] parentheses—圆括号{ } Braces—大括号Equality—等式Inequality—不等式be equal to; equals—等于[ai'dentikəl][i'kwivələnt] be identical with/to; be equivalent to—恒等于≈ be approximately equal to—近似等于< be less than—小于>be greater than—大于<< be far less than—远小于>>be far greater than—远大于['æbsəlu:t]| | Absolute value—绝对值[fæk'tɔ:riəl]n! n Factorial —n的阶乘∑ sigma [sʌ'meiʃən]the Sum of; Summation of—..的和Capital—大写的∏ the Product of --…的乘积Max/maximum value—最大值Min/minimum value—最小值Power—幂[skwɛəd]x squared—x的平方[kju:b]x cubed—x的立方Root—根['ikstrækt]Root extracting—开方5、Operation related words运算相关的词汇Addition —加法Sums —和[ə'dend]Addend —加数Solution of Addition —加法的解Addend + Addend (Augend) = Sum (Total)加数+ 加数（被加数）=和（合计）[səb'trækʃən] Subtraction 减法['minjuend]Minuend —被减数['sʌbtrəhend]Subtrahend —减数[ri'meində]Remainder or Difference 差Minuend – Subtrahend = Remainder被减数–减数= 差[.mʌltipli'keiʃən]Multiplication 乘法[.mʌltipli'kænd]Multiplicand —被乘数['mʌltəplaiə]Multiplier —乘数['prɔdəkt]Product —积Factor —因数Multiplicand × Multiplier = Product被乘数×乘数= 积[di'viʒən]Division 除法['dividend]Dividend —被除数[di'vaizə]Divisor —除数['kwəuʃənt]Quotient —商Remainder —余数[.ældʒi'breiik]Algebraic Expressions —代数式['intigrəl]Integral Expressions —整式[mɔ'nəumiəl]Monomial —单项式[.kəui'fiʃənt]Coefficient —系数degree of monomial —次数[.pɔli'nəumiəl]Polynomial —多项式[kəm'bain]Combining Similar Terms —合并同类项Removing and Adding Brackets—去和添括号Difference of Squares —平方差公式The square of the sum —完全平方公式Difference of Cubes —立方差公式[fæktərai'zeiʃən]Factorization —因式分解Abstracting Common Factors —提取公因式Fractions —分式[di'nɔmi.neitə]Denominator 分母['nju:məreitə]Numerator 分子Common factor 公因式Similar term 同类项['mʌltipl]Common multiple 公倍数[.kænse'leiʃən]Cancellation —约分lowest terms最简分式power of fractions分式的幂['rædikl]radical 根式square root 平方根cube root 立方根n-th root n次方根[.simplifi'keiʃən]Operation and Simplification of Radicals 根式的运算和化简[eks'pəunənt] Powers of Fractional Exponents 分数指数幂Multiplication for Powers with the Same Base—同底数幂的乘法Power of Exponents and Products —幂的乘方Power of Products—积的乘方Multiplication of Monomials —单项式与单项式相乘Monomials Times Polynomials —单项式与多项式相乘Multiplication of Polynomials—多项式与多项式相乘Division of Power with the Same Base —同底数幂的除法Monomials Divided by Monomials —单项式除以单项式Polynomials Divided by Monomials —多项式除以单项式6、Equation and inequality 方程和不等式Solution of Equations 方程的解Equations with One variable 单变量方程Linear Equations 线性方程[kwɑ:'drætik]Quadratic Equations 二次方程nth-degree Equations n次方程Factorization 因式分解法Completing the Square 配方法Quadratic Formula 公式求解法[dis'kriminənt]Discriminant of the Quadratic Form 根的判别式System of Equations 方程组[.saiməl'teinjəs]Simultaneous equations方程组['vɛəriəbl]Variable 变量Solution for Systems of Linear Equations with Two variables 二元线性方程组的求解[i.limi'neiʃən]Elimination by addition or subtraction 加减消元法Elimination by substitution 代入消元法['sʌbstitju:t]Substitute 代入[eks'treiniəs]extraneous root 增根[.ini'kwɔliti]First-degree Inequalities with One variable 一元一次不等式Solution for Inequalities with Absolute Values 绝对值不等式的求解['intəvəl] Interval 区间Open Interval 开区间Closed Interval 闭区间Half-Open Interval 半开半闭区间Non-ending Interval 无穷区间。

第47卷第1期Vol.47No.1计算机工程Computer Engineering2021年1月January2021融合知识图谱与注意力机制的短文本分类模型丁辰晖1，夏鸿斌1，2，刘渊1，2（1.江南大学数字媒体学院，江苏无锡214122；2.江苏省媒体设计与软件技术重点实验室，江苏无锡214122）摘要：针对短文本缺乏上下文信息导致的语义模糊问题,构建一种融合知识图谱和注意力机制的神经网络模型。

借助现有知识库获取短文本相关的概念集,以获得短文本相关先验知识,弥补短文本缺乏上下文信息的不足。

将字符向量、词向量以及短文本的概念集作为模型的输入,运用编码器-解码器模型对短文本与概念集进行编码,利用注意力机制计算每个概念权重值,减小无关噪声概念对短文本分类的影响,在此基础上通过双向门控循环单元编码短文本输入序列,获取短文本分类特征,从而更准确地进行短文本分类。

实验结果表明,该模型在AGNews、Ohsumed 和TagMyNews短文本数据集上的准确率分别达到73.95%、40.69%和63.10%,具有较好的分类能力。

关键词：短文本分类；知识图谱；自然语言处理；注意力机制；双向门控循环单元开放科学（资源服务）标志码（OSID）：中文引用格式：丁辰晖，夏鸿斌，刘渊.融合知识图谱与注意力机制的短文本分类模型［J］.计算机工程，2021，47（1）：94-100.英文引用格式：DING Chenhui，XIA Hongbin，LIU Yuan.Short text classification model combining knowledge graph and attention mechanism［J］.Computer Engineering，2021，47（1）：94-100.Short Text Classification Model Combining Knowledge Graph and Attention MechanismDING Chenhui1，XIA Hongbin1，2，LIU Yuan1，2（1.School of Digital Media，Jiangnan University，Wuxi，Jiangsu214122，China；2.Jiangsu Key Laboratory of Media Design andSoftware Technology，Wuxi，Jiangsu214122，China）【Abstract】Concerning the semantic ambiguity caused by the lack of context information，this paper proposes a neural network model，which combines knowledge graph and attention mechanism.By using the existing knowledge base to obtain the concept set related to the short text，the prior knowledge related to the short text is obtained to address the lack of context information in the short text.The character vector，word vector，and concept set of the short text are taken as the input of the model.Then the encoder-decoder model is used to encode the short text and concept set，and the attention mechanism is used to calculate the weight value of each concept to reduce the influence of unrelated noise concepts on short text classification.On this basis，a Bi-directional-Gated Recurrent Unit（Bi-GRU）is used to encode the input sequences of the short text to obtain short text classification features，so as to perform short text classification more effectively.Experimental results show that the accuracy of the model on AGNews，Ohsumed and TagMyNews short text data sets is73.95%，40.69%and63.10%，respectively，showing a good classification ability.【Key words】short text classification；knowledge graph；Natural Language Processing（NLP）；attention mechanism；Bi-directional-Gated Recurrent Unit（Bi-GRU）DOI：10.19678/j.issn.1000-3428.00567340概述近年来，随着Twitter、微博等社交网络的出现，人们可以轻松便捷地在社交平台上发布文本、图片、视频等多样化的信息，社交网络已超越传统媒体成为新的信息聚集地，并以极快的速度影响着社会的信息传播格局［1］。

黑龙江科学HEILONGJIANG SCIENCE第12卷第5期2021年3月Vol. 12Mar. 2021环境类专业“三融合"培养模式创新与实践潘法康，张瑾，唐建设,伍昌年(安徽建筑大学环境与能源工程学院，合肥230601)摘要：为了培养创新型、复合型、应用型环境类专业人才,对环境类专业"三融合”培养模式的创新与实践展开探讨。

针对高等教育同质化倾向严重、结构性矛盾突出等问题，结合近年来对环境类专业人才培养模式的实践探索，提出了学科交叉融合、产教融合、科 研教学融合的"三融合”人才培养模式，提升了人才培养质量。

环境类专业"三融合”人才培养模式的探索和实践，取得了良好的教 学成果，提升了人才培养质量，为应用型高校人才培养和教育教学改革提供了参考。

关键词：环境类专业;三融合;培养模式中图分类号：G642.0文献标志码：A 文章编号：1674-8646(2021)05 -0038 -03Innovation and Practice of Three Integration Training Mode of Environmental SpecialtiesPan Fakang, Zhang Jin , Tang Jianshe , Wu Changnian(School of Environment and Energy Engineering , Anhui Jianzhu University , Hefei 230601 , China)Abstract : In order to cultivate innovative , compound and applied environmental talents as the goal orientation , thepaper explores the “ three integration " talent training mode of environmental specialty ・ According to the problems of severe homogenization and structural conflicts of the higher education , and through combining with the practicalexploration of talent training mode of environmental specialty in recent years , the research proposes the talent trainingmode of u three integration n , including integration of interdisciplinary, industry and teaching, scientific research andteaching , in order to improve talent training quality ・ The exploration and practice of u three integration n talent training mode of environmental specialty has achieved good teaching effect , improved talent training quality , and providedreferences for talent cultivation and teaching reform of applied colleges ・Key words : Environmental specialty ； Three integration ；高校要着重培养创新型、复合型、应用型人才,为高等教育人才培养指明方向⑴。

• 1530*中华中医药杂志(原中国医药学报)2021年3月第36卷第3期CJTCM P，March 2021, Vol.36, No.3•综述•推拿治疗疼痛的研究进展及其机制探讨雷洋K2,王玉霞\周运峰uC河南中医药大学，郑州450036; 2河南中医药大学第三临床医学院，郑州450008)摘要：文章立足经典，回顾总结中医学几千年来医疗实践中推拿治疗疼痛的应用，将推拿治疗疼痛按照学科分类并结合当代医学学科中的解剖学、力学、分子生物学、神经学等学科，对推拿治疗疼痛的内在机制做尝试性探讨，从一个新的角度，即推拿临床实践结合推拿学科的发展现状，总结推拿治疗疼痛的最新进展，从临床实践应用研究中串联推拿治疗疼痛的内在机制，并且结合推拿临床实际，从人文关怀的角度以及情感引导方面丰富拓展推拿治疗疼痛的内在机制，为其进一步的研究发展提供新的思路与方法。

关键词：推拿；推拿学科；疼痛；临床研究；镇痛机制；机制探讨；进展；精神情志因素基金资助：国家自然科学基金面上项目（N o.81874513 )，河南省高等学校科研项目基础研究计划(N〇.21A360020)，河南省“中原千人计划”——中原名医（N〇.Z Y Q R201912120 )R e s e a r c h p r o g r e s s a n d m e c h a n i s m o f t u i n a i n t h e t r e a t m e n t o f p a i nL E I Y a n g12,W A N G Y u-xia1,Z H O U Y u n-feng12('Henan University of Chinese Medicine, Zhengzhou 450036, China; :The Third Clinical Medical College ofHenan University of Chinese Medicine, Zhengzhou 450008, China )Abstract: Based on the classics, this paper reviews the application of tuina on pain in traditional Chinese medical practice for thousands of years. The internal mechanism of tuina on pain is tentatively discussed according to the disciplineclassification and combining the moden medical disciplines, such as anatomy, dynamics, molecular biology, and neurology. Froma new angle, namely the combination of tuina clinical practice and development status of tuina discipline, the paper summarizesthe latest progress of tuina on pain, connects the clinical application studies to the internal mechanism, and enriches and expandsthe internal mechanism from the perspectives of humanistic care and emotion guidance combining the tuina actual clinic, in orderto provide new ideas and methods for further research and development.K e y W o rd s! Tuina; Tuina subject; Pain; Clinical research; Analgesic mechanism; Mechanism discussion; Progress;Mental and emotional factorsF u n d i n g：General Program of National Natural Science Foundation of China (No.81874513), Basic Research Plan ofScientific Research Project in Colleges and Universities of Henan Province (No.21 A360020), Henan Province 'Thousand TalentsPlan of Central Plains': Famous Doctors of Central Plains (N〇.ZYQR201912120)疼痛是多种病症所共有的一种自觉的症状(1979年，国际 疼痛研究学会（International Association for the Study of Pain，I A S P)将“疼痛”定义为“一种与实际或潜在组织损伤或与这类损伤描述相关的不愉快感觉和情感经历”。

classificationClassification is a fundamental task in machine learning and data analysis. It involves categorizing data into predefined classes or categories based on their features or characteristics. The goal of classification is to build a model that can accurately predict the class of new, unseen instances.In this document, we will explore the concept of classification, different types of classification algorithms, and their applications in various domains. We will also discuss the process of building and evaluating a classification model.I. Introduction to ClassificationA. Definition and Importance of ClassificationClassification is the process of assigning predefined labels or classes to instances based on their relevant features. It plays a vital role in numerous fields, including finance, healthcare, marketing, and customer service. By classifying data, organizations can make informed decisions, automate processes, and enhance efficiency.B. Types of Classification Problems1. Binary Classification: In binary classification, instances are classified into one of two classes. For example, spam detection, fraud detection, and sentiment analysis are binary classification problems.2. Multi-class Classification: In multi-class classification, instances are classified into more than two classes. Examples of multi-class classification problems include document categorization, image recognition, and disease diagnosis.II. Classification AlgorithmsA. Decision TreesDecision trees are widely used for classification tasks. They provide a clear and interpretable way to classify instances by creating a tree-like model. Decision trees use a set of rules based on features to make decisions, leading down different branches until a leaf node (class label) is reached. Some popular decision tree algorithms include C4.5, CART, and Random Forest.B. Naive BayesNaive Bayes is a probabilistic classification algorithm based on Bayes' theorem. It assumes that the features are statistically independent of each other, despite the simplifying assumption, which often doesn't hold in the realworld. Naive Bayes is known for its simplicity and efficiency and works well in text classification and spam filtering.C. Support Vector MachinesSupport Vector Machines (SVMs) are powerful classification algorithms that find the optimal hyperplane in high-dimensional space to separate instances into different classes. SVMs are good at dealing with linear and non-linear classification problems. They have applications in image recognition, hand-written digit recognition, and text categorization.D. K-Nearest Neighbors (KNN)K-Nearest Neighbors is a simple yet effective classification algorithm. It classifies an instance based on its k nearest neighbors in the training set. KNN is a non-parametric algorithm, meaning it does not assume any specific distribution of the data. It has applications in recommendation systems and pattern recognition.E. Artificial Neural Networks (ANN)Artificial Neural Networks are inspired by the biological structure of the human brain. They consist of interconnected nodes (neurons) organized in layers. ANN algorithms, such asMultilayer Perceptron and Convolutional Neural Networks, have achieved remarkable success in various classification tasks, including image recognition, speech recognition, and natural language processing.III. Building a Classification ModelA. Data PreprocessingBefore implementing a classification algorithm, data preprocessing is necessary. This step involves cleaning the data, handling missing values, and encoding categorical variables. It may also include feature scaling and dimensionality reduction techniques like Principal Component Analysis (PCA).B. Training and TestingTo build a classification model, a labeled dataset is divided into a training set and a testing set. The training set is used to fit the model on the data, while the testing set is used to evaluate the performance of the model. Cross-validation techniques like k-fold cross-validation can be used to obtain more accurate estimates of the model's performance.C. Evaluation MetricsSeveral metrics can be used to evaluate the performance of a classification model. Accuracy, precision, recall, and F1-score are commonly used metrics. Additionally, ROC curves and AUC (Area Under Curve) can assess the model's performance across different probability thresholds.IV. Applications of ClassificationA. Spam DetectionClassification algorithms can be used to detect spam emails accurately. By training a model on a dataset of labeled spam and non-spam emails, it can learn to classify incoming emails as either spam or legitimate.B. Fraud DetectionClassification algorithms are essential in fraud detection systems. By analyzing features such as account activity, transaction patterns, and user behavior, a model can identify potentially fraudulent transactions or activities.C. Disease DiagnosisClassification algorithms can assist in disease diagnosis by analyzing patient data, including symptoms, medical history, and test results. By comparing the patient's data againsthistorical data, the model can predict the likelihood of a specific disease.D. Image RecognitionClassification algorithms, particularly deep learning algorithms like Convolutional Neural Networks (CNNs), have revolutionized image recognition tasks. They can accurately identify objects or scenes in images, enabling applications like facial recognition and autonomous driving.V. ConclusionClassification is a vital task in machine learning and data analysis. It enables us to categorize instances into different classes based on their features. By understanding different classification algorithms and their applications, organizations can make better decisions, automate processes, and gain valuable insights from their data.。

Psychology Research, June 2023, Vol. 13, No. 6, 279-285 doi:10.17265/2159-5542/2023.06.006 Employee Attrition Classification Model Based onStacking AlgorithmCHEN YanmingShantou UniversityLIN XinyuSouth China Normal UniversityZHAN KunyeShenzhen UniversityThis paper aims to build an employee attrition classification model based on the Stacking algorithm. Oversampling algorithm is applied to address the issue of data imbalance and the Randomforest feature importance ranking method is used to resolve the overfitting problem after data cleaning and preprocessing. Then, different algorithms are used to establish classification models as control experiments, and R-squared indicators are used to compare. Finally, the Stacking algorithm is used to establish the final classification model. This model has practical and significant implications for both human resource management and employee attrition analysis.Keywords: employee attrition, classification model, machine learning, ensemble learning, oversampling algorithm, Randomforest, stacking algorithmIntroductionEmployee attrition is an important research topic in human resource management, and many scholars have attempted to establish employee attrition classification models using machine learning methods, such as application research of decision tree algorithm in talent attrition of logistics enterprises (Yang & Li, 2017), employee attrition prediction based on database knowledge discovery (Wu, 2019), research on countermeasures for the problem of core employee attrition in state-owned manufacturing enterprises (Qian, 2021), and the application of machine learning in the field of human resource management (Huang, 2022). However, these widely used algorithms have not been able to improve the accuracy of classification models.This paper uses the stacking algorithm to integrate multiple models and establish the final employee attrition classification model, which achieves an accuracy close to 1.0 and performs better than other ensemble learning algorithms in the control experiment. This model can be applied to the field of human resource management to help managers better classify employee attrition.CHEN Yanming, Bachelor degree in Economics, Finance, Department of Applied Economics, Shantou University, Shantou, China.LIN Xinyu, Bachelor degree in Management, Human Resource Management, Department of Management, South China Normal University, Guangzhou, China.ZHAN Kunye, Bachelor of Management, Management Science and Engineering, Department of Management Science, Shenzhen University, Shenzhen, China.DA VID PUBLISHINGDEMPLOYEE ATTRITION CLASSIFICATION MODEL280Theoretical FoundationMachine learning achieves specific tasks by allowing computers to learn from data and automatically adjust algorithms (Jalil, Hwang, & Dawi, 2019). Machine learning algorithms can classify, predict, cluster, or optimize based on the features and goals of the data.Ensemble learning is an algorithm that combines multiple basic machine learning models into a more powerful model. It can improve the accuracy and robustness of the model by weighting or voting the output of multiple models. Ensemble learning can improve the generalization ability of the model by reducing variance and bias.Materials and MethodsDataset Used in the StudyTo investigate and identify the significant factors responsible for employee attrition, and develop a model for classifying employee attrition, this paper employs a dataset created by IBM data scientists available on . This dataset comprises 1,470 observations, each of which represents whether an employee has resigned or not, and also includes various information about the employee. However, this dataset is considered an imbalanced dataset, because it only consists of 237 resignations and 1,233 non-resignations.Based on the information of the employee, an employee attrition classification model can be established. There are 32 variables in this dataset. We eliminated some variables that were not pertinent to our research, leaving us with 30 variables that were ultimately used. Afterward, the variable named “Attrition” serves as the dependent variable, and the other 29 variables are used as independent variables. The independent variables in this dataset contain both numerical and categorical variables, and the distribution of the numerical and categorical variables is shown in Table 1.Table 1The Proportion of Two Categories of VariablesType Number ProportionCategorical 6 20.69%Numerical 23 79.31%The partial information of numerical and categorical variables is shown in Tables 2 and 3, respectively. Table 2Partial Information About Numerical VariablesCount Min Max MeanAge 1,470 18.00 60.00 36.92Daily rate 1,470 102.00 1499.00 802.49Distance from home 1,470 1.00 29.00 9.19Education 1,470 1.00 5.00 2.91Satisfaction 1,470 1.00 4.00 2.72 Performance rating 1,470 3.00 4.00 3.15Total working years 1,470 0.00 40.00 11.28Job involvement 1,470 1.00 4.00 2.73Job level 1,470 1.00 5.00 2.06EMPLOYEE ATTRITION CLASSIFICATION MODEL 281Table 3Rudimentary Information About Categorical VariablesCount Unique Top FreqBusiness travel 1,470 3 Travel rarely 1043Department 1,470 3 Research & development 961Education field 1,470 6 Life sciences 606Gender 1,470 2 Male 882Job role 1,470 9 Sales executive 326Marital status 1,470 3 Married 673MethodsIn this paper, variables are first divided into numerical variables and categorical variables. After scaling the numerical variables and transforming the categorical variables into dummy variables, an oversampling algorithm was employed to address the issue of data imbalance. Subsequently, feature selection was performed using the Point-biserial algorithm and Randomforest feature importance ranking method. Finally, a Stacking algorithm was utilized to integrate multiple models and establish the ultimate employee attrition classification model.Data cleaning and preprocessing. The dataset under consideration is devoid of missing values and outliers. In order to mitigate the issue of overfitting, this paper utilizes the “min-max rescaling”algorithm to scale numerical variables between 0 and 1. The formula for “min-max rescaling” is as follows Equation (1):X′=X−XminXmax−Xmin(1)where X is the original value, X max and X min are the maximum and minimum values, and X' is the transformed feature value.In this paper, all categorical variables are converted into dummy variables using one-hot encoding, which are 0-1 variables.Oversampling algorithm. In this dataset, the dependent variable “Attrition” is represented by “1” for those who have left resigned and “0” for those who have not. However, the proportion of “1” and “0” is severely imbalanced, which may lead to significant errors if a classification model is directly established. In order to address this issue, this paper has adopted the random oversampling algorithm, which is the quickest and simplest method (Wang & Liu, 2020). The data before and after processing is illustrated in Figure 1.Another commonly used algorithm is SMOTE oversampling. However, through comparison, we found that the random oversampling method yields better results for this dataset. This is due to the fact that the SMOTE algorithm can not effectively address the data distribution issue in imbalanced datasets, which may result in marginalization of the data distribution and increase the difficulty of classification algorithms to accurately classify the data.Point-biserial algorithm for feature analysis. This paper conducts Point-biserial correlation analysis on all variables and the target variable “attrition”. In Point-biserial correlation analysis, the correlation represents the relationship between variables, while the p-value indicates the significance level. Variables with a p-value more than 0.05 are usually considered to have no significant correlation with the target variable. The variables with a p-value more than 0.05 are shown in Table 4.EMPLOYEE ATTRITION CLASSIFICATION MODEL282Figure 1. The data before and after oversampling.Table 4The Variables With a p-Value More Than 0.05 in Point-Biserial AnalysisVariables Correlation p-valueManufacturing director 0.030 0.137Performance rating 0.016 0.409Percent salary hike -0.014 0.472Research scientist 0.010 0.617Sales executive 0.006 0.734Hourly rate -0.003 0.843Randomforest feature importance ranking for feature filtering. The point-biserial correlation algorithm can only be employed for preliminary analysis of the correlation between each feature and employee attrition. However, when building a classification model, it is imperative to also consider the intercorrelation among features. Therefore, this paper employs the Randomforest feature importance ranking technique to select the required features for modeling purposes (Wu & Zhang, 2021).The specific process is as follows: Initially, the dataset is divided into training and testing sets in a 7:3 ratio. Subsequently, all processed variables are inputted into a Randomforest classification model. Then, a Randomforest feature importance ranking can be generated, where the sum of the importance of all features is equal to 1. Finally, features with a feature importance of less than 0.005 are filtered out. This process is shown in Figure 2.The partial feature importance ranking generated by the Randomforest classification model is shown inTable 5.EMPLOYEE ATTRITION CLASSIFICATION MODEL 283Figure 2. Process of feature filtering.Table 5The Top Eight Features of the Adaboosting Feature Importance RankingFeature ImportanceMonthly income 0.068Age 0.063Monthly rate 0.051Years at company 0.048Distance from home 0.047Total working years 0.045Percent salary hike 0.042Satisfaction 0.041Model building. The stacking algorithm is a non-linear ensemble process, which involves performing cross-validation (K-fold validation) on each base learner (first-layer model), and training a meta-learner (second-layer model) using the results of the base learners as features (Ni, Tang, & Wang, 2022). Typically, a relatively simple model is selected as the second-layer model. The main process of Stacking algorithm is shown in Figure 3.Figure 3. The main process of Stacking algorithm.EMPLOYEE ATTRITION CLASSIFICATION MODEL284This paper employs random forest and Adaboosting as the control experiments, which are two ensemble learning algorithms. Then, we employ a stacking ensemble approach, with random forest classification model, KNN model, Adaboosting classification model, decision tree model, extreme decision tree model, and logistic regression model as the first-layer models. For the second-layer model, we select a decision tree model.Experiments & ResultsExperiment EnvironmentThe dataset comes from a public database named . This experiment was done in python 3.8.0, and the configuration of the computer is shown in Table 6.Table 6The Configuration of the ComputerHardware Hardware modelCPU Intel core i7 CPU 2.90 GHZRAM 40.0 GBExperiments and ResultsFirstly, comparative experiments are conducted using random forest classification model and Adaboosting model. The random forest classification model is based on decision tree as the base model, and we select the support vector machine (SVM) model as the base model for the Adaboosting model. The results are shown in Table 7.Table 7Experimental Results of Two Classification ModelsTraining set (accuracy) Testing set (accuracy) Testing set (F1-score)Randomforest 1.0 0.985 0.984Adaboosting 1.0 0.982 0.982 Secondly, we conducted experiments using the Stacking algorithm, and the result is shown in Table 8.Table 8Experimental Result of the Stacking AlgorithmTraining set (accuracy) Testing set (accuracy) Testing set (F1-score)The ROC curve of the stacking model is shown in Figure 4.EMPLOYEE ATTRITION CLASSIFICATION MODEL285Figure 4. The ROC curve of the stacking model.From the experimental results, the stacking algorithm can improve the accuracy and F1 score and reduce the risk of overfitting to a certain extent on this dataset.ConclusionsIn this paper, we use oversampling algorithm to address the issue of data imbalance, and employe Stacking algorithm to establish the final employee attrition classification model. Through experimental comparisons, we find that it demonstrated better performance on both the training set and the testing set. However, there are still some flaws. When using the Stacking algorithm, due to the numerous models involved, the process of tuning the model parameters may become more challenging when dealing with more complex datasets.ReferencesHuang, H. L. (2022). The application of machine learning in the field of human resource management. Human ResourcesDevelopment, 23, 92-93.Jalil, N. A., Hwang, H. J., & Dawi, N. M. (2019). Machines learning trends, perspectives and prospects in education sector. InProceedings of the 2019 3rd international conference on education and multimedia technology (pp. 201-205). doi:10.1145/3345120.3345147Ni, P., Tang, K., & Wang, Z. Y. (2022). Research on sentinel-1 sea ice classification based on stacking integrated machine learningmethod. Mine Surveying, 1, 70-77.Qian, J. (2021). Research on countermeasures for the problem of core employee attrition in state-owned manufacturing enterprises.China Market, 32, 19-21.Wang, D., & Liu, Y. (2020). Denoise-based over-sampling for imbalanced data classification. In Proceedings of 2020 19thinternational symposium on distributed computing and applications for business engineering and science (DCABES 2020). Wu, D. (2019). Employee attrition prediction based on database knowledge discovery. Science and Technology & Innovation, 14,16-19.Wu, W. J., & Zhang, J. X. (2021). Feature selection algorithm of random forest based on fusion of classification information andits application. Computer Engineering and Applications, 57, 147-156.Yang, J., & Li, Y. H. (2017). Application research of decision tree algorithm in talent attrition of logistics enterprises. LogisticsEngineering and Management, 8, 154-156.。

AuctionOnline retailers Working Council for (Case Studies)(Case Studies)CIOs (1999)Awareness level Marketing % or orders correctly Sales and marketing expendituresfulfilledexpenses%of click through % of orders delivered to Attreaction of media Reliable deliverycorrect address# of referrals# of people told by one customer# of trucks% of documents used by # of fulfillment centersknowledge workers available on-line% of employees accessing Intranet at least daily Answer time Out-of-stock positions Order confirmation cycle System capacity# of orders processed time# of transactions per day # of transactions per day % of products that are # of users in live daybuilt-to-orderauctions (capacity)Logistics capacityCash conversion ratio Inventory turns/year Inventory levelsBid-to-cash cycle time Ability to handle additional traffic Network uptimeAverage time to load a page4 day delivery (partner)Revenues from # of partnersaffiliates program Logistics capacity (outsourced)Revenue breakdown by Advertising, research productand marketing # of page impression revenuesAdvertising revenues Subscription fees Revenue growthRevenue growth Value of goods traded # of products sold Administration costsOperating expenses Net assets needed to Investments support $1 worth of Cost structure outputNet profit/loss Operating profit/loss Free cash flow Gross profit marginNet profit/lossWorking capital Return on invested capitalfinancingMarket capitalization Share priceShare priceNet proceeds of IPOTable 3. Measures for E-Business Companies (continued )BrandingResource/AssestsActivities/ProcessPartner NetworkCostProfit(3) I n f r a s t r u c t u r e M a n a g e m e n t(4) F i n a n c i a l A s p e c t sRevenuedefine particular conditions of each company.。