影響不動產價格的因素，可劃分為一般因素、區域因素與個別因素三個層級，公部門評價中地價區段的劃設，於性質上即為區域因素相似、價格相近之空間同質性區域。由於缺乏理論之支持，地價區段劃設長期以來均由地價人員依經驗主觀判斷；以往雖有研究提出改進意見，然仍未能有效形成具體措施。本文基於不動產估價中影響價格因素三層級的觀點，以排除一般因素、個別因素後的迴歸殘差推計區域價格後，結合克利金法與地理資訊系統進行同質性地價區段的劃設，並將劃設之結果以變數納入價格模型中；藉由空間價值模型解釋力的顯著提昇，驗證應用克利金法輔助地價區段劃設之有效性。
Factors influencing real estate value could be classified into three categories: general, locational and property specific. As part of the public assessment process, government agencies define land value districts based on properties with similar prices and homogeneous locational factors. The delineation of these areas, however, is very subjective and lacks theoretical support. This study proposes a new approach to determining land value districts. We apply the Kriging estimation technique with GIS to the residual property value, after accounting for general and property specific factors, to define the boundary of these districts and incorporate them into the pricing model. Empirical analyses demonstrate that applying the Kriging technique in the process of delineating land value districts significantly improves the effectiveness of spatial regression models.
