数据挖掘技术在寿险代理人激励系统中的应用

发布时间：2018-04-21 14:02

本文选题：寿险代理人 + 激励方式　；参考：《湖南大学》2014年硕士论文

【摘要】：寿险作为保险行业的重要分支，是目前数据挖掘商业应用的热点领域。利用数据挖掘技术对寿险数据进行分析挖掘具有重要的现实意义。随着寿险市场的开放、外资公司的介入，竞争日趋灼热化。寿险保险公司普遍缺乏对代理人激励系统、活动以及措施的信息反馈和效果分析。因此，通过数据挖掘方法对寿险代理人激励事件、激励反馈等信息进行科学的分析研究，是提升寿险公司竞争力的重要途径。现阶段寿险公司对寿险代理人激励方式的选择上存在盲目性和不及时性，，同时对各项激励决策的收益分析不够充分。本文运用多种数据挖掘方法与人寿保险公司激励方式相结合，解决了激励理论与员工激励决策结合的问题，并对寿险代理人激励收益进行详细评估。本文重点对决策树和聚类算法展开研究，主要工作概括如下：激励方式决策时需要考虑的因素众多，如果只依据简单的人为经验进行决策将导致片面化，而通过精算分析过程繁琐并耗费大量时间。因此，本文提出基于决策树的寿险代理人激励方式决策模型，对于寿险公司代理人数据进行周密分析处理，根据设计的激励事件提取方法提取出每个代理人的激励事件，利用C4.5和Random Tree决策树预测模型，并评价分析两种决策树方法在寿险代理人数据环境下性能的差异，以得到每个代理人在自身条件下激励方式的最优决策策略。同时进行案例实证分析，利用该模型进行预测和检验，与实际精算决策结果作对比，本文方法决策F-Measure可达86.6%。基于激励方式的决策结果，本文构建了寿险代理激励方式绩效指标的聚类分析指标体系，选择相关指标数据，进而对各个聚类下激励方式分布情况进行分析探讨。通过K-Means聚类和Hierarchical聚类方法，对比分析它们在寿险公司绩效分类下的结果，从而得到当前环境下优质寿险分公司的激励方式最优比例分配方案。经案例分析证明，本文方法可为寿险公司调整各项激励方式所占比例提供有效参考。
[Abstract]:As an important branch of insurance industry, life insurance is a hot field of data mining commercial application. It is of great practical significance to analyze and mine life insurance data by using data mining technology. With the opening of the life insurance market and the intervention of foreign companies, the competition is becoming more and more hot. Life insurance companies generally lack information feedback and effect analysis on agent incentive systems, activities and measures. Therefore, it is an important way to improve the competitiveness of life insurance companies to scientifically analyze and study the information of life insurance agents' incentive events and incentive feedback through data mining methods. At present, there is blindness and intimeliness in the choice of life insurance agent's incentive mode in life insurance company, and at the same time, the income analysis of every incentive decision is not enough. In this paper, a combination of multiple data mining methods and life insurance incentive methods is used to solve the problem of the combination of incentive theory and employee incentive decision, and to evaluate the incentive income of life insurance agents in detail. This paper focuses on the decision tree and clustering algorithm, the main work is summarized as follows: There are many factors that need to be considered in the decision of incentive mode. If the decision is based on simple human experience, it will lead to one-sided, and the actuarial analysis process is cumbersome and takes a lot of time. Therefore, this paper puts forward a decision model of life insurance agent incentive mode based on decision tree. The data of life insurance company agent is carefully analyzed and processed, and the incentive events of each agent are extracted according to the designed incentive event extraction method. Using C4.5 and Random Tree decision tree prediction model, and evaluating and analyzing the difference of performance of two decision tree methods in the data environment of life insurance agent, we can get the optimal decision strategy of each agent's incentive mode under their own condition. At the same time, the empirical case analysis is carried out, and the model is used to predict and test, and compared with the actual actuarial decision results, the F-Measure of this method can reach 86.6. Based on the decision result of incentive mode, this paper constructs the cluster analysis index system of life insurance agent incentive mode performance index, selects the relevant index data, and then analyzes the distribution of incentive mode under each cluster. By means of K-Means clustering and Hierarchical clustering, the results of performance classification of life insurance companies are compared and analyzed, and the optimal incentive scheme of premium life insurance branches is obtained in the current environment. Case study shows that this method can provide an effective reference for life insurance companies to adjust the proportion of incentives.
【学位授予单位】：湖南大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：F842.62;TP311.13

【参考文献】