摘要
行政管理人员的工作的有效性决定了企业管理水平,为了提高企业管理质量,提出基于BP神经网络的行政管理人员绩效评估方法。利用改进的德尔菲法成立评估小组,初步选取评估指标,通过专家咨询的方式分析指标重要程度,采用专家积极性、权威性、集中性与离散性等标准判断指标是否科学合理,去除不合理内容,根据德、能、勤、绩、廉五项内容建立评估指标体系;利用BP神经网络的正向与反向传播性质,获取训练目标整体误差函数;将评估指标作为网络输入,确定实际输出,计算实际输出与理想输出之间的误差,经过权值与阈值修正将误差控制在最小,输出评估结果。仿真实验表明,该方法评估出的绩效结果与实际情况相符,为企业人员岗位决策提供宝贵意见。
The effective work of administrative personnel determines the level of enterprise management. In order to improve the quality of enterprise management, a performance evaluation method for administrative personnel based on BP neural network is proposed. The improved Delphi method is used to establish an evaluation team, initially select evaluation indicators, analyze the importance of the indicators through expert consultation. Standards such as expert enthusiasm, authority, concentration and dispersion are used to determine whether the indicators are scientific and reasonable, and remove unreasonable content. An evaluation index system is stablished based on the five contents of morality, ability, diligence, performance, and integrity;use the forward and back propagation properties of BP neural network to obtain the overall error function of the training target;use the evaluation index as the network input to determine the actual output. Calculate the error between the actual output and the ideal output, control the error to a minimum through weight and threshold correction, and output the evaluation result. Simulation experiments show that the performance results evaluated by this method are consistent with the actual situation, and provide valuable opinions for corporate personnel to make job decisions.
作者
李海宝
LI Haibao(College of Politics and Law,Yulin University,Yulin 719000,China)
出处
《微型电脑应用》
2023年第2期75-79,共5页
Microcomputer Applications
基金
榆林学院高层次人才科研启动基金项目(12GK26)。
关键词
BP神经网络
行政管理人员
绩效评估
指标体系
误差修正
BP neural network
administrative staff
performance evaluation
index system
error correction
