데이터 마이닝을 이용한 주택가격 결정요인에 관한 연구

전해정(Chun, Hae Jung) 한국주거환경학회 2017 주거환경(한국주거환경학회논문집) Vol.15 No.3

This study empirically analyzed the determinants of housing price using decision tree analysis, multiple regression analysis, and neural network analysis. The dependent variables are apartment real transaction price, independent variables are apartment size, apartment floor, construction year, city hall straight distance, 2 lane straight distance, 4 lane straight distance, subway station straight distance, Gangnam area, Gangseo area, urban area and Gangbuk area. As a result, the mean absolute error (MAE) of the neural network analysis is the smallest, and the linear correlation value shows the highest correlation between the real and predicted values. According to the decision tree analysis, the small size is divided into 70m2 instead of 59m2, so it can be seen that the small and medium area is popular because 2-3 person households are popular. As a result of multiple regression analysis, apartment size, number of floors, 2 lane straight distance, and 4 lane straight distance were positively affected by housing price, while city hall straight distance, subway station straight distance, population density and region from Gangnam area to Gangseo area, Gangbuk area and urban area were negative.

패널 VAR모형을 이용한 한국 주택 매매, 전세시장에 관한 연구

전해정(Chun, Hae-Jung) 한국주거환경학회 2015 주거환경(한국주거환경학회논문집) Vol.13 No.2

In this study, the dynamic relationships between the housing sales market and the Chonsei market use VAR model, and for the spatial range, 16 metropolitan cities·provinces are set, and for the timely range,from the first quarter of 2006 to third quarter of 2014 are set to divide before and after the global financial crisis to perform the empirical analysis. For the macroeconomic variables effecting the housing market, interest rate was set as CD interest rate, the economic growth and housing demand were set as number of economically active people, and finally, the housing supply was set as land transaction volume. As a result of the empirical analysis, the number of economically active people showed to have plus (+) effect and the land transaction volume to have minus (-) effect to the housing sales price in general. When we look at the difference before and after the financial crisis, in the period before the crisis, the Chonsei price and the CD interest rate had minus (-) effect on the housing sales price, but on the other hand, in the period after the crisis, it showed to have a plus (+) effect. This is because in the period before the financial cris, the housing sales market was in the increasing period, and realization of capital gain was possible through sales to have almost no influence of the lease market of housing Chonsei market to the housing sales market, but in the period after the crisis, the housing sales market maintained downward stabilization, but the Chonsei price increase to be judged that the Chonsei market has big influence on the sales market. Regarding influence of CD interest rate on the housing sales market, the reason that it showed plus (+) influence after the financial crisis is judged to be due to the reason that because the government executed continuous decrease in interest rate to activate the housing sales market, but the sales market is maintaining downward stabilization.

시계열 분석을 이용한 서울시 주택시장과 오피스시장의 상관관계에 대한 연구

전해정(Hae Jung Chun) 서울연구원 2013 서울도시연구 Vol.14 No.3

경제성장과 소득 불균형 간의 관계에 관한 연구

전해정(Hae Jung Chun) 서울연구원 2014 서울도시연구 Vol.15 No.2

저출산, 고령화가 주택가격에 미치는 영향에 관한 연구

전해정(Chun, Hae-Jung) 한국주거환경학회 2015 주거환경(한국주거환경학회논문집) Vol.13 No.4

This study examines the effects of low birthrate and aging populations on housing prices by setting proxy variables and examined the variables as a panel data across districts. Low birthrate is proxied through youth dependency ratio and number of newborns, aging through old dependency ratio, liquidity through mortgage loans, housing supply through housing permit performance, economic growth and income growth through unemployment rate, and finally, housing prices through housing transaction price index. Time frame of the data analysis was set from 1Q of 2010 through 4Q of 2013, and spatial scope was set to 25 districts within Seoul. Empirical analysis results show that youth dependency ratio and number of births has a positive impact (+) on housing prices, while old dependency ratio has a negative impact (-). 1% decrease in youth dependency ratio approximately translated to 3.18% decrease in housing prices, and 1% increase in old dependency ratio effected housing prices by a decrease of approximately 1.18%. Additionally, housing permit performance has a positive effect (+) and mortgage loans has a negative effect (-) which were different from the expected results. The study determines that this was one of the root causes of decline in housing prices during the global financial crisis.

주택 매매, 전세가격 간의 영향력에 관한 연구

전해정(Hae jung Chun) 한국부동산학회 2017 不動産學報 Vol.71 No.-

본 연구는 2008년 9월부터 2016년 9월까지의 서울, 인천, 경기도의 자료를 이용해 패널 연립방정식으로 주택매매, 전세가격 간의 영향에 관한 실증분석을 하였다. 내생변수는 주택매매가격과 주택전세가격으로 하였으며 외생변수는 주택담보대출금, 미분양주택수, 경제활동인구수, 실업률과 CD금리로 설정하였다. 주택매매가격 모형에 대한 고정효과 분석결과, 주택전세가격이 1% 상승할 때 주택매매가격은 1.309% 상승한 반면 미분양주택이 1% 상승할 때 주택매매가격은 0.014% 하락한 것으로 나타났으며 CD금리가 1% 상승할 때 주택매매가격은 0.007% 하락한 것으로 나타났다. 주택전세가격 모형 분석결과, 주택매매가격이 1% 상승하면 주택전세가격이 0.345% 상승하는 것으로 나타났으며 실업률이 1%상승할 때 주택전세가격은 0.001% 하락한 것으로 나타났다. 시뮬레이션을 한 결과, 지역별로 일부 차이는 보였으나 주택매매가격과 주택전세가격 모두 예측값과 실제값의 추이가 대체로 일치하는 것으로 나타났다. 1. CONTENTS (1) RESEARCH OBJECTIVES This study is an empirical analysis of the effects of housing price and Chonsei price on Seoul, Incheon and Gyeonggi-do from September 2008 to September 2016 using the panel simultaneous equations. (2) RESEARCH METHOD The endogenous variables were housing sales price and Chonsei price, and exogenous variables were set as mortgage loans, unsold housing, number of economically active population, unemployment rate and CD interest rate. (3) RESEARCH FINDINGS As a result of the analysis of the housing price model, the housing price increased by 1.309% when the Chonsei price increased by 1%, while the housing price decreased by 0.014%, 0.007% when the unsold housing and CD interest rate were respectively increased by 1%. According to the Chonsei price model, when housing price increased by 1%, Chonsei price increased by 0.345%, and when unemployment rate by 1%, housing price decreased by 0.001%. 2. RESULTS As a result of the simulation, it was found that predicted and actual value of housing price and Chonsei price approximately coincided with each other, though there were some differences by region.

오피스텔 매매가격 결정요인에 관한 실증연구

전해정(Chun, Hae-Jung) 한국주거환경학회 2015 주거환경(한국주거환경학회논문집) Vol.13 No.3

This study made an empirical analysis on the decision factors of the sale price of officetel located in downtown area in Seoul, Gangnam, Mapo, and Yeouido from December 2002 to February 2015 using vector auto regression(VAR). A proxy variable of economic growth was set to economically active population, interest rate to CD interest rate, and rent was set to regional annual returns of rent. As a result of the analysis, it was identified that there is difference by region but annual rent returns of officetel makes positive effect on the sale price of officetel, and it was verified that the sale price of officetel makes consistent positive effect on the annual rent returns. It was identified that the CD interest rate makes slightly positive response to the sale price of officetel, and economically active population makes positive effect on the sale price of officetel. As there is great possibility for officetel to be used as residence due to the rapid increase of single-member households and two-member households, the government authorities recognized officetel as housing that there is a need to monitor the rent and sale price constantly as the housing market.

딥 러닝을 이용한 주택가격 예측에 관한 연구

전해정(Chun, Hae Jung),양혜선(Yang, Hye Seon) 한국주거환경학회 2019 주거환경(한국주거환경학회논문집) Vol.17 No.2

The purpose of this study is to estimate housing prices using deep running. The simple RNN, LSTM, and GRU models, which are evaluated to be suitable for time series forecasting, are based on the time series data of apartment real price index, interest rate, household loan, building permit area and consumer price index. As a result of the empirical analysis, it is confirmed that the prediction power of the GRU model is superior to that of the learning data by evaluating the performance of forecasting power on apartment real price index based on the RMSE value. On the other hand, in the verification data, it is confirmed that the prediction power of the RNN model is excellent. Also, if the performance of the deep running model is evaluated with accuracy, the accuracy of the RNN model and the GRU model is the highest. As a result of this study, the government needs to build and develop a system that can predict and diagnose the housing market by using the deep learning technique that combines artificial neural network and big data to advance the housing market.

소비심리와 주택매매가격이 경매낙찰가율에 미치는 영향

전해정(Chun, Hae Jung) 한국주거환경학회 2018 주거환경(한국주거환경학회논문집) Vol.16 No.3

The purpose of this study is to analyze the effect of consumer sentiment and housing sales price on the housing auction price ratio. This study was analyzed using panel vector error correction model(VECM). The independent variable is the housing auction price ratio and the dependent variables are the housing consumer sentiment index and housing sales price. The temporal range is from July 2011 to May 2017, and the spatial ranges are Seoul, Gyeonggi-do and Incheon. As a result of variance decomposition analysis, the explanatory power of the housing auction price ratio was the largest, and the housing consumer sentiment index and the housing sales price showed the highest explanatory power. As a result of impulse response analysis, the housing auction price ratio showed a positive (+) response to the housing consumer sentiment index. The housing auction price ratio showed a positive (+) response to the housing sales price, and the housing sales price showed a positive (+) response to the housing auction price ratio. Housing sales price and consumer sentiment index showed a positive(+) response on each other. The housing sales market and the auction market were found to be coupled. Especially, it was found that consumer sentiment had a great influence on the housing market and aution market.

공간계량경제모형을 이용한 주택가격의 동학적 특성에 관한 연구

전해정(Hae jung Chun) 韓國不動産學會 2015 不動産學報 Vol.61 No.-

본 연구는 공간계량경제모형을 이용해 주택시장의 동학적 특성을 살펴보기 위해 주택매매가격, 전세가격, 주택매매거래량, 경제활동인구, 실업자수, 토지거래량과 회사채수익률로 공간패널을 구축해 실증분석하였다. 공간적 범위는 서울시 25개구로, 시간적 범위는 2006년 1분기부터 2014년 3분기까지로, 내용적 범위는 주택 중 아파트로 설정하였다. 일반회귀모형(OLS), 공간자기회귀모형(SAR), 공간오차모형(SEM)과 일반공간모형(SAC)을 추정한 결과 설명변수의 공간적 자기상관과 오차항의 공간적 자기상관을 모두 고려한 일반공간모형이 가장 적합한 모형으로 나타났다.일반공간모형 추정결과, 토지거래량이 1% 증가할 경우 주택매매가격은 0.01% 증가하고 주택전세가격이 1% 증가할 경우 주택매매가격은 0.37% 높아지고 경제활동인구가 1% 증가할 경우 주택매매가격은 0.69% 상승하는 것으로 나타났다. 통계적 유의성은 없지만 주택매매가격에 주택매매거래량과 실업자수는 음(-)의 영향을 회사채수익률은 양(+)의 영향을 미치는 것으로 나타났다. 또한, 공간상관계수가 유의수준이내에서 유의한 것으로 나타나 한 지역에서의 주택가격의 변화는 다른 지역에 영향을 미치는 것으로 나타났다. 1. CONTENTS (1) RESEARCH OBJECTIVES The purpose of this study is to examine the dynamic characteristics of the housing market. (2) RESEARCH METHOD In this study, to examine the dynamic characteristics of the housing market, the spatial panel was built with the housing price, Chonsei price, house trading volume, economically active population, number of unemployed, land trading volume and the return of corporate bond and was analyzed empirically. The spatial range was set to 25 districts in Seoul and the temporal range was established from the first quarter of 2006 to the third quarter of 2014. (3) RESEARCH FINDINGS In the results of analyzing the estimation for the general regression model (OLS), the spatially autoregressive model (SAR), the spatial errors model (SEM) and the spatial lagged term with spatially correlated error structure model (SAC), the spatial lagged term with spatially correlated error structure model (SAC) was represented as the most appropriate model. In the results of estimating SAC, it was represented that when the land trading volume was increased by 1%, the housing price was increased by 0.01%, when Chonsei price was increased by 1%, the housing price was increased by 0.37% and when the economically active population was increased by 1%, the housing price was increased by 0.69%. Although there was no statistical significance, it was represented that the house trading volume and the number of unemployed had negative (-) influence and the return of corporate bond has positive (+) influence on the housing price. 2. RESULTS Since the current raise in Chonsei price may cause the raise in the housing price, the government should understand it and needs to establish and implement diverse policies such as supplying the rented housings, etc. to stabilize the housing problem for the middle and working classes.