A Research On Quantitative Forecasting Methods Of Mobile Users

Forecast is not only the basis for decision-making, but also the tool helping to make decision. Forecasting methods and the accuracy of forecast are the core issues of the decision-support system. However, in order to make a scientific decision, accurate forecast is the necessary premise.This dissertation mainly focuses on discussion and research of forecasting models. In order to meet the characteristics and needs of forecast which will be done for mobile telecom users, the dissertation proposes a new effective method, which aims to increase the forecasting accuracy. The main research is as follows:(1) It makes an overview of basic forecasting principles and traditional prediction model. It also focuses on the analysis and discussion on causal and extension methods which are used in quantitative forecasting model.(2) In connection with mobile user development characters this dissertation does research on forecasting methods. Then it proposes a practical forecasting process and algorithm for mobile users, and summarizes the methods of improving the accuracy of forecast, which mainly include: improving the quality of basic data, the proper use of forecasting methods, combining qualitative and quantitative methods, and improving forecasting accuracy with combined models. With real examples this dissertation also gives a detailed exposition of the process and algorithm mentioned in the dissertation.(3) With the combination of the traditional statistical forecast and the control action of financial indicators, based on concept of experience departures and its adjustment, this dissertation introduces the idea of using index system as predication corrector to affect the forecast, and builds a quantitative forecasting method which fuses index verification and statistical modeling—index verification forecast method. After taking a mobile users forecasting procedure, with the model established on real case the forecast results show that this method can effectively improve the accuracy of forecast, and can greatly enhance the rationality of the forecasting system.