| In the Internet era,news reports,with their reliability and timeliness,have always been an important source of information for investors.By its report and analysis,the news media can affect market information,investor sentiment and even the investor's view of the market.Research on the sentiment of news reports helps us understand the evolution of market sentiment.In this paper,we explore the real estate market,which is strongly affected by the investor sentiment but relatively less well researched.By collecting about 200 thousand real estate news reports during 2004-2016 and using text mining methods,we constructed a new housing market sentiment index.By correlation analysis,regression analysis,time series analysis and other methods,it is proved that this sentiment index is reliable,has a strong correlation with the main indicators in the real estate market,and greatly improves the accuracy of the house price regression model.After further subdividing the housing market,it is found that the correlation coefficient between investor sentiment and house price in the first and second tier cities is significantly stronger than that of the third tier cities,and the investor's sentiment in the third tier cities is generally lower than that of the first and second tier cities.And by comparing the new house market and second-hand house market we find that there is a two-way interaction relationship between market sentiment and the real estate market price fluctuations,in which the market sentiment on the new commercial housing market has a lag effect on the house price,while this lag effect does not exist in the second-hand housing market.The main contribution of this this paper is by combining behavioral finance and big data technology,we provide a new and direct way to measure market sentiment.This method helps to reduce the errors caused by indirect indicators in measuring market sentiment,and helps to improve the independence and reliability. |
PDF Full Text Request