Research On Model Construction And Application Of BP Neural Network Technology In Management Talent Evaluation Of Company A

In the era of economic integration and market diversification,how to shape and intensify the core competitiveness of incorporation,how to adjust strategies in the changeable market environment quickly is very important.As all know,management talents with strong ability bring into a crucial role in winning the challenge and gaining advantages for enterprises.How to own and train excellent managers who have satisfied the requirements of corporation post setting,how to use and allocate talents reasonably and advantaged,how to make it used by enterprises and give full play to the maximum benefits,it is necessary to set up a scientific and effectively managerment assessment mechanism.However,the selection of indexs in the intrinsic talent assessment mechanism is so subjective and vague at all,resulting in the answers of assessment aren’t objective and precise insufficiently.Besides,we always choose blurring valuation method and AHP method in a evaluation activities,it brings about the trouble like subjectivity of assessment results and tedious evaluation process.In view of the above difficulties,I tried to introduce BP neural network technology(hereinafter referred to as BPNN)into the enterprise management talent evaluation activities.I made use of the strong parallel processing characteristics of BPNN,as well as the unique ability of adaptive learning and feedback adjustment,and combined with the traditional enterprise talent evaluation system,to build the talent assessment model based on BPNN technology,which can avoid the influence of the subjective reasons in talent assessmen.Therefore,in this study,I researched the common problems in the administrators evaluation activities of company A,built the BPNN talent evaluation model with the help of scientific methods,then I tested and applied it in the management talent evaluation of company A.For the sake of guaranteeing the positive and effective advance of the research project,I’ve read and combed the relevant principles and technologies of BPNN,management talents,talent evaluation and so on,which made full preparation for the next stage of study.I made an extensive investigation and research on the structure of managers and assessed the present situation of managers in company A.According to the features of company A’s management talents and related industries,I drew on the experience of expert argumentation,frequency statistics,questionnaire survey and literature analysis to screen out 15 evaluation indexes.After summing up and sorting out,a management talent assess indexes system including 15 second level assess indexes and 4 first level assess indexes are built for company A,and calculated the weight value of the above indexes with entropy weight method.Next,I used scientific methods and company A’s assess indexes system to design the number of neurons in each layer of BPNN,learning algorithm and fitting degree.Then,I trained and constructed the BPNN model suitable for company A’s management talent evaluation through the learning sample set and test sample set formed by the questionnaire survey results.During the test,the average output error of all the evaluation samples were no more than 5%.It shows that the BPNN talent evaluation model built in my study has excellent evaluation results in the managers evaluation of company A,which can successfully abate the evaluation error and raise the accuracy and validity of the evaluation.