基于BP神经网络模型的房价敏感性分析

为分析和测算房价，受供给和需求两方面相关指标的影响和敏感程度，本文采用BP神经网络，研究了供需相关指标的不确定性变动引起的房价变动趋势。研究表明：房价对银行贷款额度和国民经济增速最敏感，对当年施工面积和建材价格弱敏感，而对居民消费价格指数和人均可支配收入不敏感。对最敏感性指标，应从差别化信贷、拉动消费抑制投资入手；对弱敏感指标，应加大保障性住房的建设力度，完善土地市场体系，推进房地产税收改革，改善住房供应结构；对不敏感指标，应切实提高居民收入水平。本文根据房价对相关指标变动的敏感性程度，指导选择调控指标先后次序，用以提高管控的精度和住房市场供需要素调控的有效性，促进房地产市场健康运行。

A Sensitivity Analysis of Housing Price Fluctuation Based on BP Artificial Natural Network Model

The BP artificial natural network model is adopted to analyze and measure the housing price sen- sitivity by establishing the house supply and demand indexes in order to analyze the change trend of housing price. The research shows that the housing price is highly sensitive to the bank loans and GDP growth rate, weakly sensitive to the floor space of the completed buildings and building materials prices within the same year, and the least sensitive to consumer price index and per capita disposable income. In order to squeeze the hous- ing price, measures of differentiating credit, stimulating consumption and restraining investment policy are taken in accordance with the highly sensitive indices. In accordance with the weakly sensitive indices, the efforts to supply more affordable housing are made, land market system is improved and the real estate tax reform is pro- moted. The disposable income of residents should be enhanced in accordance with the insensitive indices. Ac- cording to the sensitivity of housing price fluctuation, the order of regulation indices should be selected, differ- ent efforts of market regulation should be made, and effective regulations and controls of housing markets should be implemented, thus creating a healthfully -run housing market.