벡터자기회귀모형을 이용한 온라인 뉴스기사와 아파트 매매가격의 동태적 관계 연구: 비정형 빅데이터를 활용한 감성분석 기법의 적용

In this study, we used the news on the real estate market in the online newspaper articles and compared the relationship between the apartment sales price and the real estate sentiment expressed in the online newspaper article as a representative, macroeconomic variable through the VAR(Vector Auto Regressive) model. After dividing the city of Seoul into five subregions and categorizing the apartment sale prices by size, the VAR model was applied to each. According to the impulse response analysis, the interest rate showed a negative response to the sales prices of the apartments. The sentiment index showed a positive response except for certain regions, and the magnitude of response was largest at lag two. In the variance decomposition analysis, the biggest factor for the sales price of apartments at the time of impact was the price variable itself, and the interest rate was more influential than the sentiment index. Yet in the Northeast, Northwest and Southeast subregions, the sentiment index was more influential than the interest rate and these regions are the areas that led the rise of the real estate price in the recent year. The result of this study shows that the sentiment index derived from the Big Data and sentiment analysis can be used as a useful parameter in analyzing the real estate market.