랜덤 포레스트 기반 공동주택 공용관리비 추산모델

Abstract

As multi-family housing has been chosen as the most popular types of housing in Korea, the maintenance of these facilities has become a key issue for the lives of the residents. The problem concerning the estimation of common maintenances cost is a sensitive issue and one that is currently receiving a lot of interest from the residents. Factors affecting this issue are complex and diverse. Thus, it is difficult to verify if the results of given estimates are adequate or whether or not the costs are high in comparison to other similar facilities. So, based on the historical data of multi-family housing, this study suggests an estimation model for arriving at common maintenance costs by utilizing the random forest method - a technique used in data mining. The multi-family housing data used in the configuration of the model was collected from certain regions in Seoul and, based on this, the random forest method was used to analyze the factors that influence common maintenance costs in order to develop an estimation model. This model is notable as it differed from traditional statistical methods by utilizing data mining -the random forest method- to build an estimation model for common maintenance costs. As a result of the case verification, the random forest based estimation model in this study is considered useful, and it is expected that more precise estimates will be gradually achieved as the data accumulates due to the nature of data mining and machine learning techniques.