기계 학습을 이용한 공동주택 가격 추정: 서울 강남구를 사례로
Estimation of the Apartment Housing Price Using the Machine Learning Methods: The Case of Gangnam-gu, Seoul
1
조회
0
다운로드
31
인용
0
참고문헌
키워드
초록
This study examines the applicability of the machine learning methods to the real estatevaluation. Gangnam-gu, Seoul is chosen as the study area, and the housing prices are estimatedusing the sales cases collected in 2016. The predictive power of the machine learning methods suchas SVM (support vector machine), Ensemble Model and DNN (Deep Neural Networks) is superior tothat of the multiple regression analysis (MRA) methods. Among the machine learning methods, thepredictability of the GBRT (Gradient Boosting Regression Tree) model is slightly superior to that ofthe others. In addition, we estimate the assessment prices by applying an assessment ratio toestimated housing prices. The assessment values estimated by the machine learning methodsreflect the actual transaction prices better than the actual assessment values do, and satisfy thetaxation equity requirements. Drawing on the machine learning methods, this study is expected tohelp improve the efficiency of mass appraisal such as the housing assessment.
Abstract
This study examines the applicability of the machine learning methods to the real estatevaluation. Gangnam-gu, Seoul is chosen as the study area, and the housing prices are estimatedusing the sales cases collected in 2016. The predictive power of the machine learning methods suchas SVM (support vector machine), Ensemble Model and DNN (Deep Neural Networks) is superior tothat of the multiple regression analysis (MRA) methods. Among the machine learning methods, thepredictability of the GBRT (Gradient Boosting Regression Tree) model is slightly superior to that ofthe others. In addition, we estimate the assessment prices by applying an assessment ratio toestimated housing prices. The assessment values estimated by the machine learning methodsreflect the actual transaction prices better than the actual assessment values do, and satisfy thetaxation equity requirements. Drawing on the machine learning methods, this study is expected tohelp improve the efficiency of mass appraisal such as the housing assessment.
참고문헌 (40)
- 부동산 가격공시에 관한 연차보고서, "국토교통부", 국토교통부, ~ (2012)
- 김경민, "기계학습을 통한 공동주택 가격결정요인 분석", 주거환경, 14(3), pp.29~40 (2016), 한국주거환경학회
- 김성진, "기업신용등급 예측을 위한 랜덤 포레스트의 응용", 산업혁신연구, 32(1), pp.187~211 (2016), 산업개발연구소[DOI]
- 남영우, "아파트시장예측을 위한 신경망분석 적용가능성에 대한 연구", 한국건설관리학회 논문집, 7(2), pp.162~170 (2006), 한국건설관리학회
- 민성욱, "딥 러닝을 이용한 주택가격 예측모형 연구", 박사, pp.2017
- 배성완, "딥 러닝을 이용한 부동산가격지수 예측", 부동산연구, 27(3), pp.71~86 (2017), 한국부동산연구원
- 서종덕, "데이터 마이닝 기법을 이용한 환율예측: GARCH와 결합된 랜덤 포레스트 모형", 산업경제연구, 29(5), pp.1607~1628 (2016), 한국산업경제학회
- 신성우, "서포트 벡터 머신을 이용한 건설업 안전보건관리비 예측 모델", 한국안전학회지, 32(1), pp.115~120 (2017), 한국안전학회
- 연규필, "표준주택공시가격 적정성 제고를 위한 기계학습적 접근", 부동산학연구, 21(2), pp.83~92 (2015), 한국부동산분석학회
- 유진은, "랜덤 포레스트: 의사결정나무의 대안으로서의 데이터 마이닝 기법", 교육평가연구, 28(2), pp.427~448 (2015), 한국교육평가학회
- 유하연, "통계모형을 이용한 서울시 아파트 매매가 예측 분석", 석사, pp.2016
- 이관제, "Statistical Machine Learning" (2017), (주)교우
- 이준용, "데이터 마이닝 적용을 통한 아파트 가격예측에 관한 연구", 국토계획, 42(4), pp.135~148 (2007), 대한국토·도시계획학회
- 이창로, "비모수 공간모형과 앙상블 학습에 기초한 단독주택가격 추정", 박사, pp.2015
- 이창로, "단독주택가격 추정을 위한 기계학습 모형의 응용", 대한지리학회지, 51(2), pp.219~233 (2016), 대한지리학회
- 임재만, "서울시 공동주택 공시가격 평가의 형평성에 관한 연구", 부동산학연구, 19(2), pp.37~56 (2013), 한국부동산분석학회
- 정화미, "신경망을 이용한 개별공시지가산정에 관한 연구", 국토계획, 36(7), pp.55~66 (2001), 대한국토도시계획학회
- 한국감정원, "IAAO 제정 과세평가 기준서Ⅰ(IAAO 제정 과세평가 기준서 한국어판 번역본)", 한국감정원, ~ (2010)
- 부동산 RESEARCH(Autumn), "허윤경", 한국감정원, 31~45 (2005)
- 홍한국, "신경망모형을 이용한 아파트 가격 모형에 관한 연구", 한국산업정보학회 학술대회, 2009(5), pp.220~226 (2009), 한국산업정보학회
- Antipov, E. A., "Mass appraisal of residential apartments: An application of Random forest for valuation and a CART-based approach for model diagnostics", Expert Systems with Applications, No, 39, pp.2012~
- Brieman, L., "Random forests", Machine learning, 45(1), pp.5~32 (2001)
- Drucker, H., "Support Vector Regression Machines", Advances in Neural Information Processing Systems, 9(NIPS 1996), pp.155~161 (1996)
- Fan, G. Z., "Determinants of House Price: A Decision Tree Approach", Urban Stuies, 43(12), pp.2301~2315 (2006)
- Friedman, J. H., "Greedy Function Approximation: A Gradient Boosting Machine", The Annals of Statistics, 29(5), pp.1189~1232 (2001)
- Glorot, X., "Understanding the difficulty of training deep feedforward neural networks", Proceedings of the Thirteenth International Conference on Artificial Intelligence and Statistics, 9, pp.249~256 (2010), PMLR
- Glorot, X., "Deep Sparse Rectier Neural Networks", Proceedings of the 14th International Conference on Artificial Intelligence and Statistics, pp.315~323 (2011)
- He, Kaiming, "Delving Deep into Rectifers: Surpassing Human-Level Performance on ImageNet Classification", The IEEE International Conference on Computer Vision (ICCV), pp.1026~1034 (2015)
- Hinton, G. E., "A Fast Learning Algorithm for Deep Belief Nets", Neural Computation, 18(7), pp.1527~1554 (2006)
- Kingma, D. P., "ADAM: A Methoe for Stochastic Optimization", The 3rd International Conference for Learning Representations, pp.1~15 (2015)
- Nair, V., "Rectified Linear Units Improve Restricted Boltzmann Machines", Proceedings of the 27 th International Conference on Machine Learning (ICML), pp.807~814 (2010)
- Nguyen, N., "Predicting Housing Value:A Comparison of Multiple Regression Analysis and Artificial Neural Networks", Journal of Real Estate, 22(3), pp.313~336 (2001)
- Smola, A. J., "A tutorial on support vector regression", Statistics and Computing, 14(3), pp.199~222 (2004)
- Tay, D. P. H., "Artificial Intelligence and the Mass Appraisal of Residential Apartments", Journal of Property Valuation & Investment, 10, pp.525~540 (1992)
- Vapnik, V., "The nature of statistical learing theory" (1996), Springer
- Vincent, P., "Extracting and composing robust features with denoising autoencoders", proceedings of the 25th international conference on Machine learning(ICML), pp.1096~1103 (2008)
- 국토교통부, "땅 위에 건물 지으면 땅값 하락", 국토교통부, ~ (2016)
- "케라스:파이썬 딥러닝 라이브러리"[URL]
- "Why are deep neural networks hard to train?"[URL]