표본 주택 가격 기반 부동산 가격지수 산정: 머신 러닝 방법의 활용을 중심으로

Estimating the Real Estate Price Index Based on Sample House Price: Focusing on the Use of Machine Learning Method

The Purpose of this study is to estimate the real estate price index based on the ‘estimated price by machine learning’. The price of a sample house was estimated using the machine learning method ‘Random forest’ and ‘Deep neural networks’, and the real estate price index was calculated using the Jevons index calculation method. First, the result of the study showed that the RF index and DNN index are similar, and the variability was changed according to the learning period. Second, the RF index and DNN index showed similar long-term trends compared to the KAB index, but it was found that there was a considerable difference in short-term trends. Third, the RF index and DNN index were found to be more variable than the KAB index, KB index, and real transaction price index, and the relationship with real transaction price index could not be confirmed. If the researcher’s qualitative analysis on the RF index and DNN index is added, it is expected that there is a high possibility of utilization as a new price index that can improve existing price index.