The Application And Research Of BP Neural Network On The Prediction Of Website Traffic

The informatization of today's society has reached a very high level, and information has become the main driving force of promoting the civilization blood flowing, information technology and information industry has in the stages of being developed rapidly. Website as an individual or corporate representative in the online world has a significant place, and site traffic, whether for site managers or site users, has a very important practical significance.As site traffic affected by various factors to varying degrees,,and the network traffic characteristics with different environments are dynamic, nonlinear process of change, the current forecasting methods and models available on the network cannot afford a full and accurate description of the flow characteristics, also can not accurately describe the site traffic of the future trends.Artificial neural networks originated in the 40s of last century, as a mathematical model of information processing based on modern biology research in human brain tissue to simulate of biological neural networks. Because of its good tolerance and adaptability, is widely used in biological neuroscience, psychology, computer science and engineering applications and information science and other fields, and have achieved good results.This paper attempts to use artificial neural network traffic prediction models and site evaluation index system of a combination of methods to forecast traffic on the site, by building a index system can be a complete description of factors that influence the web site traffic, then use of artificial neural network itself has a good self-organization, self-studying and strong nonlinear approximation ability, building a site traffic forecasting model, hoping to get satisfactory results. Main tasks of this paper are as follows: (1) Overview of the existing site traffic forecasting methods and models, collate and analyze the artificial neural network method and demonstrate its feasibility for site traffic forecasts.(2) On the basis of previous studies, constructed website traffic forecasts evaluation index system by using of SPSS statistical analysis, combined with artificial neural network build a prediction model of the website traffic, followed by an empirical research of the model taking a example of 'www.ahtv.cn'.(3) Through the analysis and summary of empirical results, combined with the problems found during the study, presented the inadequacies of this study and suggestions for improvement to provide a reference for future research.