民营经济是我国社会主义市场经济中的重要组成部分，其健康发展对国民经济的高质 量发展具有重要的影响作用。近年来，R 市 T 镇汽摩配行业中小民营企业员工离职率一 直居高不下，成为企业发展的重要阻力。特别是在当前由新冠疫情引起的复杂的社会背景 以及紧张的经济背景下，保持员工队伍的稳定对中小民营企业战胜各种风险挑战以及实现 可持续发展具有重要意义。 一直以来，员工离职倾向研究都是企业人力资源管理领域中的关注焦点。传统的人力 资源管理理论主要针对人和组织开展研究，而并未对人力资源数据做过多的分析研究。随 着信息技术和大数据技术的不断发展，企业人力资源数据分析能够为员工离职预测提供更 有影响力的决策依据。鉴于此，本文采用问卷调查方法，收集分析了 T 镇汽摩配行业的 中小民营企业人力资源数据，结合离职理论基础利用信息技术和大数据技术对离职因素的 相关性进行实证分析，得出了对离职影响较大的因素，并建立了员工离职倾向的逻辑回归 分析模型、决策树模型以及随机森林模型。 本文通过模型测评得到了最优预测模型，利用模型可测试企业员工离职倾向，针对离 职倾向较高的员工制定特定的挽留政策。根据研究结果，本文从合理构建考评体系、制定 薪酬制度、完善员工学习培训制度、 孕育企业文化建设四个方面对 T 镇汽摩配行业中 小民营企业提出了员工离职行为管理建议，为中小民营企业离职研究提供管理思路。 关键词：离职倾向；影响因素；预测模型；模型评价；浙江理工大学硕士专业学位论文 II Abstract Private economy is an important part of China's socialist market economy, its healthy development plays an important role in the high-quality development of national economy. In recent years, the turnover rate of small and medium-sized private enterprises in auto and motorcycle parts industry in T Town, R City has been high, which has become an important obstacle to the development of enterprises. Especially in the current complex social and economic background caused by coVID-19, maintaining the stability of the workforce is of great significance for small and medium-sized private enterprises to overcome various risks and challenges and achieve sustainable development. The research on employee turnover intention has always been the focus in the field of human resource management. Traditional HRM theories mainly focus on the research of people and organizations, but do not do much analysis and research on HUMAN resource data. With the continuous development of information technology and big data technology, enterprise human resource data analysis can provide more influential decision-making basis for employee turnover prediction. In view of this, this paper USES the questionnaire survey method, collect and analyze the T town for automobile industry of the small and medium-sized private enterprises human resource data, combined with the departure of theoretical basis for using information technology and large data correlation factors of departure for the empirical analysis, obtained the departure of influential factors, and establish the logistic regression analysis model of employee turnover intention, the decision tree model and stochastic model of the forest. In this paper, the optimal prediction model is obtained by model test, using the model can be used to test enterprise employee turnover intention, for higher turnover intention of employees, formulate specific retention policy, according to the results, this paper constructed appraisal system, establish compensation system, improve the system of staff training and enterprise culture construction from four aspects of T town for automobile industry of small and medium-sized private enterprise employee turnover behavior management advice, puts forward the management idea is presented for the study of the small and medium-sized private enterprises exit. Key Words: Turnover intention; Influencing factors;Prediction model;Model evaluation浙江理工大学硕士专业学位论文 III 目 录 摘要..............................................................................................................................................I Abstract......................................................................................................................................... II 目 录...........................................................................................................................................III 1 绪 论........................................................................................................................................... 1 1.1 研究背景..............................................................................................................................1 1.2 研究目的..............................................................................................................................2 1.3 研究意义..............................................................................................................................2 1.4 创新点和技术路线..............................................................................................................3 1.4.1 本文创新点..................................................................................................................3 1.4.2 本文技术路线..............................................................................................................4 1.4 本章小结..............................................................................................................................5 2 理论基础和相关研究综述.........................................................................................................6 2.1 离职相关理论性分析..........................................................................................................6 2.1.1 离职的概念..................................................................................................................6 2.1.2 离职倾向的分类..........................................................................................................6 2.1.3 离职的影响因素..........................................................................................................7 2.2 国内外离职预测的理论基础..............................................................................................7 2.3 本章小结............................................................................................................................11 3 企业人力资源数据收集及分析...............................................................................................12 3.1 R 市 T 镇汽摩配行业人力资源现状.................................................................................12 3.2 人力资源数据收集............................................................................................................13 3.3 人力资源数据集整体情况................................................................................................16 3.4 本章小结............................................................................................................................17 4 离职因素相关性分析与预测模型建立...................................................................................18 4.1 数据清洗与归一化处理....................................................................................................18 4.1.1 数据清洗概念及过程................................................................................................18 4.1.2 数据归一化概念及过程............................................................................................19 4.2 离职因素相关性分析........................................................................................................21 4.3 离职预测模型建立............................................................................................................26 4.3.1 逻辑回归分析模型....................................................................................................26 4.3.2 决策树模型................................................................................................................27 4.3.3 随机森林模型............................................................................................................28 4.4 预测模型评价....................................................................................................................30浙江理工大学硕士专业学位论文 IV 4.5 本章小结............................................................................................................................31 5 离职倾向预测及离职行为管理建议.......................................................................................33 5.1 离职倾向预测....................................................................................................................33 5.2 企业员工离职行为管理建议............................................................................................36 5.3 本章小结............................................................................................................................38 6 总结和展望............................................................................................................................... 39 6.1 研究结论与讨论.................................................................................................................39 6.2 研究不足与展望................................................................................................................39。。。。。。以下内容略