基于因子分析的特征价格模型改进

The Improvement of Hedonic Price Model Based on Factor Analysis

摘要: 特征价格模型在房地产价格模型和对非市场产品的价值评估中得到广泛应用，但是一般的特征价格模型不能有效地确定产品价值的影响因素，涉及指标众多，信息重复，使拟合的准确性降低。本文将因子分析应用于特征价格模型，用几个不相关的潜在因子替代了重复的指标变量，并且用因子旋转解释潜在因子的现实意义，从而保证了信息全面、变量简单。最终用潜在因子按照特征价格模型理论进行多元回归，消除了多重共线性，提高了回归准确性。

Abstract: Although hedonic price model is extensively used in real estate prices and the value assessment of non-market product, duplicate information along with numerous indexes which it concerned with not only makes its fitting accuracy declined but also prevents people giving a further interpretation to the influencing factors of the object’s value. This paper makes factor analysis applied to Hedonic Price Model, replaces repeating index variable by irrelevant potential factors and explains the practical significance of potential factors by factor rotation so that the features of full information and simple variable are ensured. Hedonic price model can be improved with factor score applied to multiple regression model, which contributes to a higher prediction accuracy and a better intepretation of the influencing factors of value.