Application Research of Rough-GA-BP Method in the Real Estate Early-Warning System

摘要

Early-warning system of China's real estate is still in the development of a sound stage, and there are following two main aspects. Firstly, the selection of indicators is to be improved. Secondly, predictive capability of the turning point about the real estate business cycle is to be improved. Based on the above-mentioned problems, the Rough-GA-BP model proposed is applied to the real estate early-warning system. Based on Rough-GA-BP model, the prediction is divided into the following steps. First, based on the theory of Rough-Ann we can make indicators screening. Second, the genetic algorithm (GA) is applied to BP neural network, and according to the GA-BP model, we have optimized its hidden layer structures and the initial right. Third, make a good use of Rough-GA-BP model constructed to predict the turning point of the real estate business cycle. Finally, compare the results predicted based on the Rough-GA-BP model with the ones based on the traditional methods.