공공데이터를 활용한 아파트 매매 가격 결정 모형의 예측 능력 비교 : 서울 강남구 지역을 중심으로

초록

This study compares the forecasting ability of apartment sale pricing models in Gangnam-gu, Seoul, using various machine learning algorithms for real estate public data such as apartment transaction prices from the Ministry of Land, Infrastructure and Transport. The multiple linear regression model has advantages that it is easy to understand, but it has a disadvantage that it can’t easily meet assumptions such as normality and independence of residuals. Especially, it has lower performance than machine learning algorithms in prediction ability. In this study, random forest and support vector machine algorithms were used to compare the estimation performance of apartment sale prices. As a result, the predictive power was significantly improved compared to the multiple regression model.

키워드