헤도닉가격모형과 서포트 벡터 회귀분석모형을 이용한 공업용 부동산의 가격추정 = Price Estimation of Industrial Property Based on Hedonic Price Model & Support Vector Regression Model

다국어 초록 (Multilingual Abstract)

The purpose of this study is to compare the price estimation of Hedonic Price Model(HPM) and Support Vector Regression(SVR) model for industrial property which include factory site and factorybuilding aiming to seek for any possible mutual supplementation. In the pursuit of this purpose, a linear model was selected as the optimum model for regression analysis and as the result, 11 variables were finally chosen, including Land Area, Adjacent Road Width, Office Size, Main Building Height, Years Existed, Distance from a Main Arterial Road, Distance from an Expressway, Distance from a Railway Station, Distance from an Administrative District, In or out of Daegu Area, and GRDP per person. SVR model was set based on those variables. HPM and SVR model were used for price estimation and their comparison, and the result was the SVR model being more accurate in predictive power. The following are the implication of the analysis, First, SVR model is free from the basic assumption of regression like HPM. Second, SVR model is easier to apply compare with HPM. Third, SVR model is excellent in predictive power. Fourth, the capability of SVR model in price estimation has been confirmed. Based on the results of this study, The complementation of the merit of regression analysis like HPM which is generally used in the estimation of industrial property prices and the merit of SVR model would be made good use in price estimation and prediction of industrial property.