The Research And Application Of Income Forecast Of Kunming Telecom Based On Data Mining

As is well known, the huge, numerous, complicated and disorderly history data of the Telecommunication company conceals a lot of worthy information which can help people make policy decision. With the increasingly competition of Telecommunication company in domestic market, It is an urgent matter for Telecommunication company to find out these management decision-making information.Based on the management analysis system of Kunming Telecom ,This thesis analyze and forecast the business income in Kunming Telecom by data warehouse and the analysis Technique of data mining's time series.Aiming at the topic of business income's analytical forecast, the data sampling which have something to do with each business income was extracted from the affair center of Kunming telecommunication. Then, we build data warehouse which made us to get gathered data, apply two current extensive forecast model which are not-steady time sequence but forecast efficiently — model Auto-Regressive and model ARIMA to forecast the business income of Kunming Telecommunication company .These models are checked and compared in AIC, SBC standard, remaining check of white noise , model parameter, and the results of experiment were also valuated.The putting forward of ARIMA model greatly makes people draw up not- steady time sequenceto match accuracy and it has already become the most classicalest time sequence model. The ARIMA model uses remaining method to withdraw certain information, so it isn't very ideal for direct explanations. The model of Auto-Regressive withdraws sequence information from the fixed factors which not only keep a direct view, but also make use of remaining information, the two draw up to match each other very closely, in this thesis the ARIMA model is especially introduced.This text discussed the whole process of forecast of the income, accordingly, homologous calculate way and SAS tools are used to carry on, which have certain reference value for people to do related analysis in business.