Study On The Prediciting Model Of Chibor

Interest rate, as the price of capital, is one of the most important variables on the financial markets.Interest rate is also the most basic problem on the financial markets. Moreover, Interbank Offered Rate is the core interest rate. As the trading volume of the interbank market is very huge, Interbank Offered Rate can reflect market trends better than other money market rates. With the trading volume of the interbank market surge and the volatility of Interbank Offered Rate accelerate, the implementation of monetary policy and commercial bank management are affected. Therefore, predicting Interbank Offered Rate accurately or not is a very important issue on our money market. In this paper we study on the predicting model of CHIBOR.The paper is divided into five parts:The first part describes the research background, current situation, the research innovation and limitation. At first we introduce the importance of predicting inter-bank offered rate accurately; Then we review the current literatures about predicting the inter-bank offered rate and point that most current predicting models of the inter-bank offered rate are the time series forecasting model; Finally, we project the the innovation and limitation.The second part describes the use of ARMA time series model in forecasting interbank offered rate. First we introduce the basic definition of ARMA model and then set up ARMA model modeling according to Box and Jenkins's 4 steps thought. At last we make the forecast within and outside sample with the model, and conclude:the. prediction accuracy of AMRA model within and outside the sample is high, however, there are also some disadvantages:for example lagged effect exists in the prediction process, the model does not adapt volatile external environment and the long-term prediction is fatigue.The third part describes the use of BP neural network in forecasting interbank offered rate. First we introduce the concepts and theories of neural network technology and the basic model of this chapter-BP neural network model, which includes the learning process of BP network, algorithm, and pros and cons; Then we set the parameters and samples to establish the BP neural network; Finally we use BP network to predict and conclude:BP network prediction effects excellently, especially no lagging effect exsits, but there are also some shortcomings:the forecast precision is less than ARMA model and the error of short-term forecast is volatile.The fourth part describes the use of combined model in forecasting interbank offered rate. This chapter describes the three combined models respectively:the right combination, the optimal weighted combination and the ARMA-BP model, and use the three combined models to predict interbank offered rate.