Research On Combination Weighting Method Of The Performance Evaluation Of Enterprise Informatization

This paper analyzed the characteristics of enterprise informatization and recognized the complexity of the performance evaluation of IT applications in enterprises.It is not just a result of subsequent performance, but also contains the process performance such as optimization of business operations and improvment of sustainability.But firstly,the traditional performance measurement system (financial perspective) is easy to overlook hidden benefits and long-term benefits.Secondly,the traditonal performance measurement method,whether objective or subjective weighting method,does not lead to convincing results because of limited data samples and affect of subjective factors existensing in the expert evaluation.Therefore, this paper established a comprehensive integrated performance evaluation of enterprise informatization system and model on the basis of combination weighting ideas.In this paper, the research framework is as follows:(1)It constructed a framework and an index system of enterprises informatization performance evaluation based on IT BSC (This article aimed at manufacturing industry).(2)It designed diffrent questionnaires in view of the subjective weighting method and the objective weighting method,which were distributed,collected and sorted in order to obtain valid data.(3)It selected AHP as a subjective weighting method and entrop method as an objective weighting method,and then calculated the weights.(4)It established a evaluation model of enterprise informatization performance based on the BP neural network model.This paper ultimately selected 12 evaluation indexes from the four dimensions and established IT performance evaluation index model for the manufacturing enterprise. We got the initial weights by using AHP method and entropy method,and the preliminary evaluation of enterprise IT performance which were collated into 5 clusters.And then we designed a network of 12-14-5 (ie 12 input layer neurons,14 hidden layer neurons,5 output neurons layer) of neural network. The final test samples were passed the test which has practical meaning.