[4Yin2] Interactive session 2

[4Yin2-54] Sequential Estimation of Explanation Variables for Housing Prices by Kalman Filter

〇Satoshi Yamada¹, Kazushi Okamoto¹, Atsushi Shibata² (1.The University of Electro-Communications, 2.Advanced Institute of Industrial Technology)

Keywords:Housing Price Estimation, Kalman Filter, Geographically Weighted Regression, Explainablity, State Space Model

Most previous studies have challenged to visualize housing price structures with regression modeling in order to solve information asymmetry in real estate industry. However, there are few models which deal with temporal changes of the housing price structures based on the past states. This study proposes a housing price regression model with Kalman filter which considers explanatory variables as a state. For the proposal, the prediction accuracy is measured and the temporal changes of the housing price structures are visualized by using two years of new house data in Tokyo in LIFULL HOME’S dataset. According to the experiment, it is confirmed that the proposal achieves superior prediction accuracy and processing times compared with the geographically weighted regression model.