主成分分析法在住宅特征价格模型中的应用

Application of principal component analysis in bousing hedonic price model

线性函数以计算简便、结果精确而广泛应用于特征价格模型中。但考虑线性函数中自变量数目较多并且存在一定相关性的问题,本文提出应用主成分分析法,对特征价格函数中的系数进行标准化处理,排除了变量间的线性关系对特征价格模型所产生的误差。同时,对西安市住宅市场进行实证研究,建立基于主成分分析的特征价格模型,并与线性函数的参数估计结果进行比较,以实际数据证实主成分分析法的可行性和有效性。

The linear function has already been widely applied in hedonic price models because of its simple calculation and accurate result. Considering certain problems in large number and relativity of the variables, this paper suggests applying principal component analysis （PCA） to standardize the coefficients in the hedonic pricefunction, excluding the error that the relativity of coefficients in hedonic price model produced. As an example, the housing market in Xi＇an is studied using the PCA-based hedonic price model. Compared with the parametric estimation results of linear function, it is proved that the method is useful and superior.