Application of GRNN Based on Correlation Analysis to the Real Estate Valuation
Guanchen Ding · Cehui dili xinxi · 2014
The theory of the Generalized Regression Neural Network is introduced in this paper. Then a price forecasting model is established using the real estate data of Jingtian district,Shenzhen.To reduce effect of human factors during the real estate appraisal process,a Correlation Analysis between twelve influence factors and real estate prices is calculated. And then six most significant impact factors are picked out as the network input. The GRNN with optimized network input has a better forecast result,when compared with the traditional BP neural network. And it is more conducive to the estimate analysis in the real estate markets.