基于主成分分析的房地产市场情绪指数构建 ——以重庆市为例

以构建房地产市场情绪指数为研究目标，爬取2019.07~2020.06 期间房地产舆情数据，基于情感词典及机器学习算法进行文本情感分析，构建舆情情感指数作为直接情绪指标；选择房地产开发投资额、新开工建筑面积、每月房屋成交建筑面积、楼面地价、个人住房贷款发放金额5 个指标作为间接情绪指标，结合文本情感分析技术得到的舆情情感指数，利用主成分分析方法构建复合市场情绪指数；与同期百度指数及冰山指数比较进行检验分析。结果表明，房地产市场情绪指数构建合理，可以反映房地产市场发展走势，为构建房地产市场情绪指数提供了一个合理方法。

Building of Investor Sentiment Index in Real Estate Market Based on PCA ——A Case of Chongqing

This research aims to build the real estate market sentiment index. Firstly，we crawl the real estate public opinion data ofvarious website from July 2019 to June 2020，analyze the text sentiment based on sentiment dictionary and machine learningalgorithm，and build the public opinion sentiment index as the direct sentiment index. Secondly，five indicators are selected asindirect emotion indicators，including real estate development investment，new construction area，monthly housing transactionconstruction area，floor price and individual housing loan amount. By using principal component analysis method, the five indirectemotion indicators and the public opinion emotion index obtained by text emotion analysis technology are combined to constructcomposite market emotion index. Finally，a test analysis is carried out to compare this paper’s index with Baidu Index and icebergindex in the same period. The results show that the real estate market sentiment index is constructed reasonably and can reflect thedevelopment trend of the real estate market. This paper provides a reasonable method for the construction of the real estate marketsentiment index.