The Applied Research Of Combined Model In Forecasting The Interbank Lending Rates Based On Artificial Intelligence Optimization

Interbank lending rates, as the most sensitive financial market interest rate, has a guiding role in the financial markets.In bank lending rates forecast, give priority to the short-term prediction by a day mainly. This article provided in Shanghai and Chinese interbank market on July1,2013to June30,2014the inter-row overnight interest rates, forecast the lending rates by a single model and combined model with step and multi-step prediction. In this article, we first put forward using the ARIMA model, support vector regression (SVR) model, a Bayesian neural network (Bay-NN) model, the prediction results of three basic models were analyzed. Based on this, we used the ARIMA, SVR and Bay-NN as the basic model, and the cuckoo search (CS) optimization algorithm for weight, put forward the CS-combination forecast model. The portfolio model that can combine the three basic model advantage (linear and nonlinear), make full use of each basic model of sample information, effectively reduce the single model in the limitations in the process of interest rate forecast. Finally make the combination model of predictive value in comparing the real value. The empirical results show that the combined forecasting model based on cuckoo weights optimization algorithm can effectively improve the prediction accuracy, reduce the error. The combination forecast model can effectively predict Shanghai and Chinese interbank market interbank lending rates, it is practical and scientific.