The purpose of this study is to solve the practical problem of maintaining the predictive accuracy of a pre-owned condominium price model in changing market conditions. To improve the prediction accuracy of the model, it is effective to subdivide the dataset into highly homogeneous neighborhoods. However, subdivision leads to the problem that predicted prices are updated only for regions where new transactions occur, but not for neighborhood regions. To solve this problem, we propose a method of dividing the data set into the smallest units, and then reintegrating the divided regions without degrading prediction accuracy.
