Early Warning Based on BP Neural Networks for Real Estate Market in Shenyang
Jie Kou · Journal of Northeastern University · 2010
An early warning index system was set up by combining the characteristics of the real estate market in Shenyang with such functions as monitoring,evaluation and early warning.According to the artificial neural network(BP model),11 data as early warning indexes were chosen in the range from the year 1998 to 2008,among which the data of the first 9 years were used as trained samples with the data of the last 2 years as tested samples.Trained with the Bayesian normalized network 1 538 times,the target accuracy was achieved with high dependability as verfied by the network simulation results.The comparison between modeling analysis and actual market situation showed that the market is fairly brisk in both 2007 and 2008.The model proposed will benefit the sound and sustainable development of real estate market in addition to its effect of early warning on the market.