Microblog Life-cycle Analysis And Forecasting

Recently, social network sites has been the one of many main routes from which online users get access to necessary information due to the development of information technology so that MicroBlog prevails in China in a rapid speed. It is possible that online users share with others via sending data to all kinds of clients on the internet. Numerous mainstream portal web sites have already provided with MicroBlog service which attracts enormous amount of online users. Both research and commercial value of MicroBlog have dramatically increased because of the enormous stockpile of information MicroBlog contains and the large scope that information can spread via MicroBlog. Moreover, the evaluations given by the public online toward products are of instructive assistance to the improvement of products. Therefore, Micro Blog data can be widely used in brand evaluation, the forecasting of business, intelligence gathering of Competitive products, Consumption decisions and customer relation management. However, the external influence factors(such as WeChat social platform) along with Weibo itself factors(weak relations, reducing valuable information and less profit points, etc.) have greatly hindered the further development of weibo.Thus, it is of great importance to analyse MicroBlog in terms of the life cycle. MicroBlog life cycle prediction has become a new topic. This dissertation will use the prediction method of product life cycle to predict the MicroBlog life cycle.Mathematical model is the main approach to the prediction of product life cycle. It utilizes the single index fitting of the product life cycle curve. Whereas, the single index method obviously is not adoptable in this case because of numerous Weibo index factors, the high_correlation of each index and the large amount of data. On this condition, this dissertation utilizes principal component analysis and Entropy weight/value distribution to join the index factors into one factor. This dissertation put forward a life cycle prediction algorithm based on BCG matrix and regression analysis, It improves the prediction accuracy.This research includes the following aspects:(1) Extracting the characteristics of weibo principal component, and assigning weights for the extracted characteristics of weibo to facilitate the next forecast analysis.(2) Weibo life cycle prediction model. Weibo life cycle analysis is divided into two parts, the rough judgment and curve fitting of weibo life cycle. Life cycle analysis model contains a judgement matrix to judge life cycle phase, Life cycle prediction model contains corresponding curve prototype for different life cycle stages. According to the stage of life cycle, we select different curve, using least square method to determine the curve parameters.(3) Putting forward a life cycle prediction algorithm based on BCG matrix and regression analysis.We use BCG matrix to analysis weibo's life cycle stage, using the regression analysis methed to fitting the curve of weibo life cycle stage and predicting the next stage of weibo life cycle.The experimental results show that this algorithm can solve the problem of complex indicators and high_correlation data weibo met in the life cycle prediction index. This algorithm has higher feasibility.