Research On The Early-warning Model Of Employee's Invisible Absence Behavior Based On BP Neural Network

Proliferated enterprise crisis issues caused by employee's invisible absence behavior have aroused high attention of enterprise managers.As a result,the management demands for dealing with employee's invisible absence behavior have been ever increasing in recent years.Invisible absence behavior is triggered by employees' complicated intrinsic psychological motivation and its serious consequences to enterprise are often neglected due to its implicity and uncontrollability.Thus,it is quite necessary to analyze the causalities in-depth and to construct a reasonable early-warning model of invisible absence behavior,so as to enable enterprise receive crisis signal on time,get early involved accordingly and take appropriate management measures.This is not only a kind of welfare for employees,but also a high return investment to enhance enterprise's performance and productivity.By tracking the research progress at home and aboard and deeply analyzing the problem of employee's "at work but out of it" which is widely existed in enterprise for the time being,this research redefines the concept and basic features of invisible absence behavior,systematically analyzes and refines the influencing factors of invisible absence behavior by using the relevant theories and methods of statistics,management,etc.,constructs a conceptual model of employee's invisible absence behavior with the method of structural equation model,and analyzes the reliability-validity,the path and the fitting degree of the model through SPSS 21.0 and Amos 21.0 softwares,so as to verify the correctness of the influencing factors and the conceptual model.On this basis,the research designs warning degrees by using K-means clustering algorithm and relying on the early-warning theory framework,constructs a early-warning model of employee's invisible absence behavior by taking BP neural network as the guiding ideology and utilizing the neural network toolbox of Matlab R2015 a,trains the early-warning model repeatedly through designing a simulation experiment in order to acquire the optimal state of the model and heighten the effectiveness and accuracy of early-warning.This model realizes a one-to-one correspondence between early-warning state and the countermeasures,and provides a decision-making basis for enterpris to accurately predict future risk state and to improve its management style accordingly with the risk state as the standard.From the existing literatures,scholars at home and abroad failed to place enough emphasis on the problem of invisiable absence.This research carries out a deep analysis on this problem with a purpose to dig up the connotation and external effect of the behavior itself,to enable enterprise managers to discover and prevent the problems,so as to guarantee the security and stability of the enterprise.Therefore,this dissertation provides a theoretical basis for the study of invisible absence problem,and be of an important guiding significance for the healthy development of enterprises.