摘要
笔者以美国房价为样本,选取了可能影响房价的13个因素,以特征价格法的思想为基础,运用回归分析、聚类分析、神经网络模型、回归树模型等统计工具,构造房价关于特征因素的测度模型,探究房价的关键作用因素。笔者通过分析得出,room(房间数)和lsat(低层阶级人口比例的对数)是两个最重要的变量,它的政策含义是,国家应该大力增加房产信息透明性,并在房地产政策调整过程中注意阶层差异化的影响。
In this paper,we use the housing price of Boston as the sample and select 13 factors which may affect it to construct the measurable model about the housing price and the hedonic factors,and try to find the key factor affecting the housing price based on he-donic price by using the statistical tools such as regression analysis,cluster analysis,artificial neural networks model and regression trees model. Then,we compare the models according to the predictive effect and further discuss the key factor. This study finds that the num-ber of rooms and LSAT ( logarithm of the ratio of the lower class) are the two most important variables,which provides the message to policymakers that the state should greatly increase the transparency of real estate information,and pay attention to the impacts of differ-ent classes may have on the adjustment of real estate policy.
出处
《经济经纬》
CSSCI
北大核心
2015年第2期127-132,共6页
Economic Survey
基金
国家自然科学基金重点资助项目(70932004)
关键词
房价
社会结构
回归分析
聚类分析
神经网络
Housing Price
Social Structure
Regression Analysis
Cluster Analysis
Artificial Neural Networks
