Research on the Structural Optimization of the Data Mining-Based Enterprise Human Resource Management

Abstract

Excellent human resource management can lead to the benign development of enterprises. This paper briefly introduces the application of back-propagation (BP) neural networks for human resource demand prediction in order to facilitate the structural optimization of human resources. The traditional BP neural network was improved using the particle swarm optimization algorithm. An analysis was carried out on an electric power company (company X) in Zhengzhou City, Henan province. The optimal node number and the type of the activation function in the hidden layer of the improved BP neural network were determined by an orthogonal comparison test. The results showed that the optimal number of nodes in the hidden layer of the improved BP neural network was 15, and the activation function was sigmoid. The multiple regression analysis method and the traditional BP neural network were compared with the improved BP neural network. The errors of the multiple regression analysis method in predicting the demand of 2018, 2019 and 2020 were 1.24%, 1.68% and 1.89%, respectively; the corresponding errors of the traditional BP neural network were 0.51%, 0.63% and 0.74%, respectively; the corresponding errors of the improved BP neural network were 0%, 0.10% and 0.10%, respectively. The prediction error of the improved BP neural network was the smallest in the workforce demand prediction of the same department.