서울시 주택가격지수의 모형별 예측력 비교 분석

Comparative Analysis for Predictability of Housing Price Index by Model in Seoul

1. CONTENTS (1) RESEARCH OBJECTIVES The purpose of this paper is to compare housing price index predictabilities by using ARIMA and artificial neural network models in Seoul housing market. (2) RESEARCH METHOD This paper uses analytical techniques such as ARIMA and artificial neural network models. (3) RESEARCH FINDINGS The research findings of this paper are as follows. First, the result for Seoul housing market showed that there was statistically no difference between ARIMA and artificial neural network models in housing price index predictability. Second, the results for Seoul housing sub-market showed that there was statistically no difference between ARIMA and artificial neural network models in housing price index predictability except for the fourth group sub-market. The result for the fourth group sub-market, which was relatively low in housing price index increasing rate, showed that the housing price index preditability of artificial neural network model was better than that of ARIMA model. 2. RESULTS The results of this paper may be summarized as follows: the housing price index predictability of artificial neural network model was more excellent than that of ARIMA model in the Seoul housing sub-market, which was relatively low in housing price index increasing rate.