The Application and Predictive Models Base on Bayesian Classifier in Electronic Information Industry

Abstract

This article analyzes the 120 months level of electronic information industry development data of 35 indicators in 30 provinces from the ZhongHong database. A Bayesian Classifier forecast model is built to measure the development level of electronic information industry index α, considering four core modeling properties which are smoothness, randomness, trends and normality. This model is used for simulation and forecasts of actual application of the electronic information industry. Specifically, according to the variation of index α in each region at different time, time-series forecast model is built for the forecasts of index α in each region in 2014; Bayesian classifier model is then constructed to evaluate the development level of electronic industry given the index α in each region. The Bayesian classifier can classify different α into year so that it can produce a development trend of electronic industry in each province. The results indicate that the development levels of electronic industry in Shanghai and Beijing is higher than other provinces. However, the differences between the levels in the two advanced cities and Zhejiang, Tianjin, Jiangsu, Fujian and Guangdong are shrinking from 2013 to 2014. In addition, the results also show that there are no record in the evaluation of ‘slightly advanced’ region. However, the electronic industry in Shandong, Hubei, Liaoning, Ningxia and Hainan will have good prospects in 2014, filling up the area of ‘slightly advanced’ region. This model requires not much data, which can be easily transplanted. The model is suitable for research and analysis of macro mechanism without much data and is also a reference for government department and related businesses.