Quantile Regression Analysis of Residential Mass Appraisal Models-Improvement in Atypical Housing Appraisal

Abstract

隨著國內不動產市場M型化推案趨勢，非典型住宅（如：高總價豪宅和低總價小套房）類型逐漸增多，對於此類型產品的估價精準度也需要提升。從過去研究發現，最小平方迴歸估計忽略各特徵屬性對價格條件分配的差異。本研究乃以分量迴歸方法建立住宅大量估價模型，藉以瞭解住宅特徵對於不同價格分量的差異，實證結果發現以最小平方迴歸模型估計相較於分量迴歸，對於一樓、頂樓、車位、區位等變數有高估或低估的情形。比較估值模型預測精確度，本文透過30次重複實驗，發現分量迴歸對於兩側尾端樣本有較佳的預測能力。從實證方法而言，本文改進以最小平方迴歸模型對兩尾端價格高估或低估問題；就實務應用方面，隨著不動產產品差異度增加，以及新版巴塞爾協定(Basel Ⅱ)實施對不動產價值更新的需求，分量迴歸模型可提升兩尾端估計精確度，並提供住宅大量估價系統另一種資產重估方法。

Analysis of the current domestic trend of residential types shows that high-priced and low-priced dwelling units are gaining popularity. Thus, the estimation of popularity of these two classes of residence should be made more precise. Because ordinary least square regression cannot signify the variation caused by different quantile functions of a conditional distribution, this study estimates the housing price by quantile regression. The models are compared with ordinary least square regression and quantile regression. Empirical results reveal that the distributions of some variables, such as first floor, top floor, parking lot, location, are different between the two models. These differences are easily underestimated or overestimated when applying ordinary least square regression. Results of hit rate and mean absolute percentage error based on 30 repeated experiments using random sampling indicate that quantile regression estimates more accurately than ordinary least square regression on two-tailed distribution. For mass appraisal applications, a quantile regression advances the estimate on two-tailed price, and provides a new method for asset reevaluation of banks.