서울시 주거용 건축물 경매 빈도를 활용한 주거불안 지역 예측 및 검증

This study aims to identify the variables predicting residential real estate auctions at the administrative district level in Seoul, South Korea, where housing anxiety has become a social issue due to high interest rates and charter fraud. The study forecasted areas of housing instability using time-series analysis and deep learning models. Data from 2017 to 2022 were used to predict figures for 2023 , and those from 2017 to 2023 were used to predict the figures for the first half of 2024. The results showed that the predictive accuracy of the model was higher for apartments than for non-apartments, largely because of the differences in sample size between them. However, predicting the distribution of auctions was more effective for non-apartment properties because there were significantly more cases of non-apartment auctions than those of apartment auctions. Moreover, economically vulnerable populations predominantly live in areas with a high concentration of non-apartment auctions, and housing instability is likely to worsen in these areas. This study’s findings provide key policy implications for preemptive measures for mitigating housing anxiety, such as preventing charter fraud, supporting affected tenants, and addressing the risks of involuntary relocation due to auctions.