基于主成分分析与BP网络的房地产投资风险预测

Real Estate Investment Risk Gorecasting Model Based on PCA-BP Neural Network

房地产投资具有投资金额大,回收期长,投资风险受多种因素影响等特点,通过主成分分析法将众多指标进行综合,消除样本间的信息重叠,降低BP网络的输入维数.针对房地产投资风险系统的非线性特征,运用BP网络的高度非线性映射能力,对房地产投资风险进行预测,并且与直接用BP网络方法作了对比分析,结果表明:基于主成分分析与BP网络预测结果更精确.

Real estate investment has some distinguishing features： heavy invest and long recovery etc. Its forming is influenced by various factors. indexes, eliminated information overlapp Through principal component analysis, we have synthesized numerous ing of s real estate investment risk system, using BP network altitudinal nonlinear map, we have efficiently predicted real estate investment risk. After having been compared with BP network model, the result proves that the method is of high precision and great applicability.