透天住宅格局因子與性價比之研究

The Research on Configuration Factors and Price-Performance Ratio of Street Houses

鄭吉盛(CHI-SHENG CHENG)

指導教授 ： 陳清山

朝陽科技大學/設計學院/建築系建築及都市設計碩士班/碩士(2020年)

摘要

透天住宅是台灣城鄉常見的住宅類型，許多縣市仍是以透天住宅為主流，由內政部營建署統計104年住宅狀況抽樣調查，顯示透天住宅占住宅49.20%的比例(內政部營建署)，可見透天住宅為台灣主要的住宅型態之一，尤其受到中南部民眾的青睞。
本研究蒐集全台透天住宅案例作為支持向量機模型訓練，為得知透天住宅格局評分、單位造價及性價比，將蒐集之眾多透天住宅案例交由專家評分，並建立資料庫，經由灰色理論進行分析及灰關聯度的計算，最後透過支持向量機學習專家們的評分模式。
首先，利用主成份分析法將51項透天住宅格局影響因子，歸納為16項主成份，並運用灰色理論探討獨棟、雙拼及連棟三種型態之透天住宅格局主成份與參考序列的灰關聯度。可從研究結果得知，與三項參考序列關聯性較高之主成份不盡相同，格局評分中，以(主臥衛浴)、(空間數量)及(臥室長寬比)較為重要；單位造價中，以(對外開口長度)、(B臥室附陽台)及(主臥室長寬比)較為重要；性價比中，以(客廳長寬比)、(B臥室附陽台)及(C臥室)較為重要，表示上述主成份於透天住宅有著較高的影響力。
本研究亦運用支持向量機與多元迴歸，推論獨棟、雙拼、連棟透天住宅之格局評分、單位造價及性價比。研究結果得知，支持向量機相較於多元迴歸結果更適合用於推論透天住宅之格局評分、單位造價及性價比。本研究結果有利於往後設計者及建築師在設計透天住宅格局時，能夠透過支持向量機的推論，快速得到設計者的性能評價。

關鍵字

透天住宅 ； 性價比 ； 主成份分析法 ； 灰色理論 ； 支持向量機 ； 多元迴歸

並列摘要

Street houses are a common type of housing in Taiwan. Street houses are still the mainstream in many cities. The Construction and Planning Agency Ministry of the Interior executed a sample survey of housing conditions in 2015, which showed that street houses accounted for 49.20% of the housing. It can be seen that street houses are one of the important residential types in Taiwan, especially favored by southerners.

This study collected all street house cases as support vector machine model training. In order to know the street house configuration score, unit cost and price-performance ratio, this study collected many street house cases scored by experts, and established a database, used the grey theory to calculate the degree of grey relationship. Finally, the support vector machine was adopted to deduce the scoring model of the experts.

First, using principal component analysis to summarize 51 street house influential factors into 16 principal components, and using grey theory to explore the degree of grey relationships between the 16 principal components and the reference sequences. It can be seen from the research results that the main components with high correlation with the three reference sequences are not the same. In the configuration score, (master bedroom bathroom), (number of spaces) and (bedrooms length to width ratio) are more important; as unit cost, (external opening length), (second bedroom balcony) and (master bedrooms length to width ratio) are more important; in price-performance ratio, (living rooms length to width ratio), (second bedroom balcony) and (third bedroom) is more important, which means that the above-mentioned main components have a higher influence on street houses.

This study also used support vector machine and multiple regression analysis to infer the configuration score, unit cost and price-performance ratio of three-type street houses. The research results show that the support vector machine is more suitable for inferring the street house's configuration score, unit cost and price-performance ratio than multiple regression analysis. The results of this study are expected to benefit designers and architects in designing the configuration of street houses in the future, through the support vector machine inference model, and quickly get the designer's performance evaluation.

並列關鍵字

Street Houses ； Price-Performance Ratio ； Principal Component Analysis ； Grey Theory ； Support Vector Machine ； Multiple Regression Analysis

參考文獻

中文參考文獻

(1) 鄧聚龍，1982，《灰色控制系統》，華中理工大學出版社，pp.24-55，中國。

(2) 趙志豪，2003，"高雄市高層大坪數住宅平面設計者所依循準則之調查研究"，碩士論文，逢甲大學建築及都市計畫研究所。

(3) 張曉平，2004，"室內環境之美學因素與美感反應關係之研究：以住宅客廳為例"，碩士論文，中原大學室內設計研究所。

(4) 陳南宏，2005，"運用類神經網路於住宅空間格局之相似度判斷"，碩士論文，國立臺灣科技大學建築系。