1
3
E89020
3
006
結合不連續選擇理論及模糊集合論對住宅消費決策行為之研究
A Study of Housing Choice Behavior by Jointing with Discrete Choice Theory and Fuzzy Set Theory
1.陳彥仲
1.Yen-jong Chen
1.國立成功大學都市計劃學系
1.Department of Urban Planing, National Cheng-Kung University
001,002,003,004
1.國立成功大學都市計劃學系
1.Department of Urban Planing, National Cheng-Kung University
002
1.行政院國家科學委員會
1.National Science Council
A.14 計畫執行期間(起):1999-08-01
A.14 計畫執行期間(訖):2000-07-31
2000-01-042000-03-16
A.16 收到日期:2000-09-20
2
1998年1月至1999年12月在「大台南都會區」興建完工的透天及大樓新屋個案中已完成交易且房屋為自有的家戶

1

672


672
001,003
01
C.2 聯絡日期:2000-10-13
2
1.國立成功大學都市計劃學系
1.Department of Urban Planing, National Cheng-Kung University
1
C.7 資料公開日期:2000-08-01
2
008
2
1
1
1
2
3
3
001
1,2
1.Logit模型
2.住宅消費
3.解模糊化
4.模糊Logit模型
5.模糊集合論
6.模糊語意尺度
1.Defuzzification
2.Fuzzy Linguistic Scale (FLS)
3.Fuzzy Logit Model (FLM)
4.Fuzzy Set Theory
5.Housing Consumption
6.Logit Model
一 般家戶在選擇住宅購屋時,包含甚為複雜的決策過程,其中的不確定性甚高。如何衡量這些複雜的因素,使其成為適切的購屋評估決策準則,為本研究所欲探討的主 要課題。本研究針對家戶購屋決策歷程,以消費者行為理論為基礎,應用改善後的模糊語意尺度法(FuzzyLinguistic Scale, FLS)來萃取家戶重要購屋決策屬性變數,以提出家戶購屋決策評估指標的合理思考架構。進而將其結合不連續選擇理論中的Logit模型(logit model),以發展出新的「模糊Logit模型」(FuzzyLogit Model, FLM)。並結合相關分析與因子分析來修正模糊Logit模型,以分別校估及建立新的家戶購屋決策模式。最後將上述三種模型參數校估結果進行比較分析。在 實證分析上,本研究選取台南都會區剛購買新屋的家戶作為研究對象。 研究結果指出,透天部分產生購屋前重視程度的四種填答類型與購屋後滿意程度的五種填答類型。大樓部分得到購屋前重視程度與購屋後滿意程度均各有六種填答類 型,且其無論在得到之語意尺度型態及實際可分析樣本數上均較透天部分為多。就各模糊語意變數的含意可發現,無論大樓或透天受訪家戶對購屋評估準則之重視程 度大都傾向較為負面的語意措辭,而購後滿意度亦產生類似較為負面的看法。 其次在模型之比較上,無論在透天或大樓住宅區位選擇模型方面,模糊多項Logit模型的配適能力、預測準確率、重要解釋變數的顯著性及參數係數符號等方面 皆較原始多項 Logit模型來的佳,顯示在模型的整體解釋能力甚至資訊處理上不僅能充分改進原始多項Logit模型之缺點,亦較符合人類真實認知感受與實際行為的決策 過程。

In this lstudy, we proposed to estimate the housing choice probability according to the well-developed Logit models jointed with the fuzzy linguistic scale (FLS) method. The choice indexes are defined by using the FLS method and then applied to the Logit models. The choice behavior potentially can be designed as a simultaneously choice structure (known as the multinomial Logit model). The new model named as \"Fuzzy Logit Model\"(FLM) is proposed to be more capable to deal with the problem of qualitative variables in a quantitative-based approach. The accuracy of the new model will be examined by comparing the results from the original models. As a result, we come to the following conclusions: (1) We derive six different linguistic types and five different semantic transformations in single unit housing, and eight different linguistic types in multiple unit housing. The definitions of various types of fuzzy number are different and unsymmetrical. The opinions of popular people intend to more negative linguistic terms whether the scores of fuzzy number in the important degree before housing purchase and the satisfied degree after housing purchase or not. (2) The results of case study of decision-making behavior for housing purchase in Tainan urban area clearly support the hypothesis and show that modified Fuzzy Multinomial Logit (FMNL) model would sufficiently improve the shortcomings of traditional Logit Multinomial model (MNL). And it is also closer to human realistic cognition and the procedure of decision-making behavior in housing consumption.

D.16 完成檢誤日期:2002-01-16
D.17 預定釋出日期:2002-07-04
D.18 初次釋出日期:2002-07-25
D.19_1 最新版釋出日期:2002-07-25
1
1
1
2
1,3
10.6141/TW-SRDA-E89020-1
https://efenci.srda.sinica.edu.tw/webview/index.jsp?object=https://efenci.srda.sinica.edu.tw:80/obj/fStudy/E89020
追蹤清單
下載(0)
申請(0)
遠距(0)