次市場與樓層效應：不動產估價中異質性特徵的分析Submarkets and Floor Height Effects: Analyzing Heterogeneity in Real Estate Valuation

本文提出新的次市場劃分法，結合尺度地理加權迴歸(Multiscale Geographically Weighted Regression, MGWR)與集群分析法，並以此為基礎估算住宅大樓的樓層效用比，進行房價的水平與垂直構面分析。以台南市住宅大樓的交易資料分析的結果顯示，住房特徵在地理空間對房價具顯著的異質性影響。此外，樓層效用比架構在不同次市場間亦存在差異，多數次市場呈現樓層越高，房價越高的正斜率曲線，然而中樓層的住宅在特定區域更受青睞。本文提供房價在水平面與垂直面異質性的證據，對不動產估價之理論支持，也對銀行、購屋者、建商及政府機構掌握房市動態及制定政策方面具參考價值。

關鍵詞：次市場、地理加權迴歸模型、自然分類法、K平均數法、樓層別效用比

This study proposes a new method for delineating submarkets by integrating a local spatial econometric model (MGWR) with cluster analysis to define submarkets and analyze floor utility ratio, thereby providing a comprehensive analysis of both the horizontal and vertical dimensions of housing prices. Using transaction data from high-rise residential buildings in Tainan City, this study reveals significant spatial heterogeneity in the horizontal and vertical of the housing market. The findings further indicate variations in the vertical structure of buildings across submarkets, with most submarkets exhibiting a positive slope, where higher floors command higher prices. However, in certain areas, the floor effect curve shows an initial rise followed by a decline, suggesting a non-linear relationship. This study offers evidence of price heterogeneity in housing market, underscoring the necessity of incorporating the horizontal and vertical dimensions into valuation models. The findings not only provide theoretical support for improving the accuracy of real estate valuations but also offer valuable insights for banks, homebuyers, developers, and government institutions in understanding housing market dynamics and informing policy decisions.

Key words: Sub-markets, Geographically Weighted Regression, Natural Breaks, K-means, Floor Utility Ratio