%0 Journal Article
%T 基于机器学习的员工流失预测<br>Employee Attrition Prediction Based on Ma-chine Learning
%A 李思晴
%A 罗鄂湘
%J Modeling and Simulation
%P 5531-5542
%@ 2324-870X
%D 2023
%I Hans Publishing
%R 10.12677/MOS.2023.126502
%X 员工流失是当今组织中的重要问题，通过机器学习等技术对员工离职进行事前预测，有助于提升企业人力资源管理的前瞻性。本文首先从采集的数据集中提取有用的且适合模型训练条件的数据，进行数据清洗和探索性分析，了解各特征分布情况；使用One-Hot编码和标签编码相结合的方式进行编码，然后采用RF-RFE方法对数据集中的特征进行筛选后进入分类模型；为保证模型预测的准确性，采用了五种不同的机器学习算法，包括SVM、DT、RF、LightGBM和WRF，来建立模型对员工流失情况进行预测。综合结果显示，LightGBM算法在预测性能方面表现出色，其准确率达到了0.87；进而通过SHAP输出特征重要性排名，发现城市发展指数、工作经验和培训时数等是影响离职的重要因素，可以为企业员工保留和后续人才招聘决策提供技术支持。<br />
Employee turnover is an important issue in today’s organizations. Predicting employee turnover in advance through machine learning and other technologies can help improve the foresight of enter-prise human resource management. In this paper, the useful data are extracted from the collected data set, and the data cleaning and exploratory analysis are carried out to understand the distribu-tion of features. A combination of One-Hot coding and label coding was used, and then the features in the dataset were screened by RF-RFE method and entered into the classification model. In order to ensure the accuracy of the model prediction, five different machine learning algorithms, includ-ing SVM, DT, RF, LightGBM and WRF, were used to build the model to predict the employee turnover situation. The results show that LightGBM algorithm has excellent prediction performance, and its accuracy rate reaches 0.87. By outputting SHAP feature importance ranking, it is found that urban development index, work experience, training hours, etc., are important factors affecting turnover, which can provide technical support for employee retention and follow-up talent recruitment deci-sions.
%K 员工流失预测，机器学习，RF-RFE，LightGBM，WRF<br>Employee Attrition Prediction
%K Machine Learning
%K RF-RFE
%K LightGBM
%K WRF
%U http://www.hanspub.org/journal/PaperInformation.aspx?PaperID=75867