The Research Of Evaluation Model About Web Service Based On BP Neural Network

Along with the electronic commerce rapid rise and the rapid development of Web applications, the Web service becomes an effective solution of the electronic commerce and complex communication service for business. The cyberization and globalization of Web applications and business as well as the automation of information processing and information integration, provide a broad application background and market for development and research of web service. The Web service has been the development of large scale, integrated. In the process of technological development, the tendency is more and more complex. The platform development commercial companies constantly improve their own R&D products. As Web service market continues to expand, inevitably bring about a new round of technological competition, two big platform developers camp has made a white-hot stage competition, many enterprises are facing a difficult choice, J2EE or NET. The two camps have undoubtedly has great strength, this paper aims to analyze different aspects of the practicality of products designed by the two camp. In order to offer some references in academic field and technical field.This thesis is the premise to construct the electronic commerce, qualitative and quantitative evaluation on the two mainstream Web service providers camp, J2EE and. NET.The first step, with the combination of quantitative and qualitative methods put forward the assumption.first of all focus on the architecture, the function etc of two platforms, make an analysis of J2EE and NET advantages and disadvantages. Through the questionnaire survey data, to use of SPSS survey data for reliability analysis, the analysis reaches reliability index, further on existing data were analyzed by AHP consistency analysis, to determine the weight of each index, at last combination of the theoretical value and the weights to present the hypothesis that the practical of J2EE is more than the practical of.NET.The second step, normalized the questionnaire survey data, get the training samples and test samples, and then using BP neural network method of training samples, and finally obtains the evaluation, through the final evaluation results verify the hypothesis is correct before.