Research And Practice On The Deposit Forecasting Model Of The Housing Provident Fund

In the city of Jixi, a large quantity of funds has been billed into the housing provident fund management Center. Questions as how to better manage those funds with modern methods and in compliance with relevant state policies so as to make full use of the funds have also arisen before Jixi and the other housing provident fund management centers in this country.With the housing system reform, people pay more attention to the housing provident fund. The amount of the housing provident fund is increasing, and data related to the fund is expanding. How to successfully identify the information useful for fund management and how to analyze the information by means of quantitative methodology and model are two questions important for the more scientific and rational policy-making in housing provident fund management.The housing provident fund deposits involve in the collection, withdrawal, payment and return of individual housing loans, and the purchase quota of treasury bonds. The collection of housing provident fund and payment of housing loans are the two main tasks of the housing provident fund management, the fulfillment of which bears great effects on the success of national housing reform policies. Therefore, it is of great significance to analyze housing provident fund data through forecasting methods, by which the amount of deposit funds could be anticipated for the next segment and the highest possible amount of funds could be used to pay the loans.In this paper, the author tries to look into the economical forecasting method and its features, compares the time series of ARIMA forecast model with artificial neural network. works out a linear regression forecast model on the basis of the data collected from the payment and return of housing provident fund loans in 2003, and the ARIMA forecasting model and GRNN neural network forecast model based on the data got from the fund collection and withdrawal since 2000. Then, comparison is further made between the results of the above two models to explore an effective application of self regression model and artificial neural network to the management of housing provident fund.On the basis of the forecast model and the existing management system of housing provident fund in Jixi, a new forecasting system of housing provident fund management has been designed and practiced in the city. Now the system has been partly adopted in the management center and proved effective in improving the rationality and standardization of housing provident fund management.