Research On Liquidity Risk Early Warning Combination Model Based On GA Algorithm

On the background of the global economic integration, the rise of the Internet financial trend as well as the process of our financial industry opening to the outside world moving forward, our country’s commercial Banks are facing great opportunities and challenges. In recent years,liquidity risk as one of the main risks faced by commercial Banks, more and more cause the attention of people. So accurately in advance to understand that the bank liquidity level and early warning of commercial Banks liquidity risk level becomes more important. The purpose of this paper is developing a set of effective and accurate method for early warning of commercial Banks liquidity risk that has strong practical significance.As the commercial Banks liquidity risk early warning problem, on the basis of domestic and foreign research on liquidity risk early warning model, according to the principle of non-linear combination forecast and genetic algorithm (GA) to optimize model portfolio weight coefficient method, the construction of commercial bank liquidity risk early warning based on GA algorithm network combination model. First of all, using the thought of the non-linear combination forecast and the principle of genetic algorithm (GA), established the basic ideas of the combination of genetic algorithm to optimize model weights. Then, the BP network model and RBF network model as the combination model of single model, build commercial bank liquidity risk early warning based on GA algorithm network combination model, and the combination model of prediction error is the sum of absolute value as a fitness function. Finally, based on the BP network model, RBF network model, and based on the combination of GA algorithm applied the results of comparative analysis. GA combined model in the prediction of approximation effect, the robustness of the model and the level of liquidity risk on the classified accuracy has more advantages, it can be seen that combination model can effectively combine the advantages of single model, can be more effective implementation of commercial Banks liquidity risk early warning.