数据挖掘技术在人力资源需求预测中的应用研究

发布时间：2018-12-31 08:18

【摘要】：随着社会的不断发展和进步,企业的人力资源数据越来越多。如何从这些人力资源数据中挖掘出信息,已经成为企业发展建设过程中的迫切需求,而数据挖掘技术为企业人力资源的需求预测带来了新的机遇。通过良好的人力资源挖掘及预测技术可以帮助企业管理者清晰的了解企业人力资源状况,为其提供正确的人力资源需求预测数据,以帮助管理者更好的管理企业。本文首先对需求预测的定性与定量分析进行研究,对数据挖掘算法进行研究并分析对比,在深入了解预测挖掘技术的基础上,确定多元线性回归方法,通过建模、统计指标分析及显著性检验得到回归方程,并依此对企业人员数进行需求预测,验证多元回归模型的可行性。通过研究传统的标准BP神经网络算法的优点、流程及存在的不足,提出一种对BP神经网络的学习率进行改进的新方法,该方法采用差异化的学习率,实现不同节点之间权重的动态调节,从而加快权重的调整速度,提高训练的收敛速度。本文详细描述了学习率自适应BP神经网络算法改进流程及推倒过程,详细分析了算法的优化步骤及流程,并进行实例验证。最后完成了人力资源需求预测系统的设计与实现,通过数据库设计及模块设计,主要实现了人力资源管理和人力资源预测两个功能模块,通过对人力资源数据进行管理及需求预测,得出结果,并最终实现预测结果的可视化。
[Abstract]:With the development and progress of the society, more and more human resource data are available. How to extract information from these human resource data has become an urgent need in the process of enterprise development and construction, and data mining technology has brought a new opportunity for enterprise human resource demand prediction. Through the good human resource mining and forecasting technology, it can help the enterprise managers to understand the human resource situation clearly, provide the correct human resource demand forecast data for them, in order to help the managers to manage the enterprise better. In this paper, the qualitative and quantitative analysis of demand forecasting is studied, and the algorithms of data mining are analyzed and compared. On the basis of in-depth understanding of prediction mining technology, the multiple linear regression method is determined. The regression equation was obtained by statistical index analysis and significance test. According to this, the demand of enterprise personnel was predicted, and the feasibility of multivariate regression model was verified. By studying the advantages, flow and shortcomings of the traditional standard BP neural network algorithm, a new method to improve the learning rate of BP neural network is proposed. The dynamic adjustment of weights between different nodes is realized so as to speed up the adjustment of weights and improve the convergence speed of training. This paper describes in detail the process of improving and pushing down the learning rate adaptive BP neural network algorithm, analyzes the optimization steps and flow of the algorithm in detail, and verifies the algorithm with an example. Finally, the design and implementation of human resource demand forecasting system are completed. Through database design and module design, two functional modules of human resource management and human resource prediction are realized. By managing the human resource data and forecasting the demand, the result is obtained, and finally the visualization of the forecast result is realized.
【学位授予单位】：华北电力大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：F272.92;TP311.13

【参考文献】