Research On Inter-bank Loan Interest Rate Prediction Method Based On SVM

In recent years, with China's market-oriented interest rate continuing to accelerate the process, tools of the interest rate of China's macro-control were increasingly used. The interest rate risks of the commercial banks were more apparent. Inter bank market as banks and other financial institutions for temporary, short-term lending market funds, is an important part of money market. The scale of their transactions of Inter bank market is increasing. With more and more frequent changes in inter-bank offered rate, for the commercial banks, the major players of the inter bank market,the volatility risk can not be underestimated. Therefore, accurate forecasts trends inter-bank offered rate for commercial bank lending to guard against interest rate risk is significant.In this paper, a forecast of the new model of inter-bank offered rate based on support vector machine model was made. First of all, the domestic and foreign interest rates forecast was introduced. And then with the interest rate decision from the Marxist theory, the combination of Western economics theory of interest rate decisions, some of the factors that influence interest rates were explored. And then from various angles the impact of interest rate changes in China's economic variables, net RMB deposits and loans, the exchange rate targets, money supply, fixed-asset investment, inflation, the stock market turnover, the difference between import and export of several variables was selected. Then starting with the risk of VC dimension and expectations of both the theoretical basis of SVM elaborate, the principle of SVM (including linear and nonlinear regression) was focusing on. Adjusting the model from the number of support vector, the kernel function was improved with the generalization ability in the model and accuracy to find a balance between. Finally, China's inter-bank offered rate for the forecast simulation. So SVM-based inter-bank offered rate of time-series experiments and the impact of factors that inter-bank offered rate with the selection of economic variables was carried. Through a comparative analysis of the two, a suitable prediction model was found. In the end, a systematic framework for inter-bank interest rate projections was made, and was perfected.Based on the prediction analysis, forecasting short-term interest rate (about 4 months) prediction with time series accuracy relatively higher than that of ARIMA method has better prediction accuracy, however, long time forecast error variance are larger; proceeding with the interest rate on the interest rate factors to predict long-term predictions better fit. So the inter-bank offered rate can be better predicted through a combination of both in an actual forecast.