不動產評價模式特徵價格法與逼近調整法之比較

On the Comparison between Hedonic Regression and Grid-Adjustment Techniques in Real Estate Appraisal

特徵價格法主要根據不動底特徵為成套組合（package）之特性，利用不能加以分割出售的特徵組合做為衡量不動產價格的重要因素；逼近調整法則是市場比較法與特徵價格法的結合，運用市場比較法的過程，並利用特徵價格估計值作為調整的依據，以排除人為主觀判斷的缺失。本研究目的即藉著特徵價格法及逼近調整法來估計不動產之價值，並利用台灣資料將兩者績效作一比較。實證結果顯示：（1）逼近迴歸調整法的評價，績效平均而言略優於特徵價格迴歸法；（2）各種特徵函數模型中，對數線性模型優於其他二函數模型──線性函數與指數函數。（3）不同區域因為發展的特性不同，主要影響不動產價格的因子亦可能不同，因此其特徵函數可能不同。

The purpose of this paper is to compare the performance of two real estate pricing techniques, hedonic regression and grid-adjustment techniques. The grid-adjustment technique is a combination of the market method and the hedonic regression method. It integrates the hedonic regression into the adjustment process of the market method to reduce the barely unavoidable subjectivity during the adjustment process. The results of our investigation are as follows: (1) Overall speaking, grid-adjustment approach is superior to hedonic regression approach in terms of the predicting performance; (2) The log-linear functional form is the most appropriate for hedonic equation among linear, exponential and log-linear forms. (3) Because of the differences in the development of different areas, the major factors influencing real estate price may be different. Hence, the specification of hedonic equations may be also different in different areas.