為改善不連續選擇理論中的多項Logit模式之IIA特性問題及便於進行模式間之差異比較,以及為檢驗住宅消費類型選擇與住宅單元選擇(例如有無社區管理)之間是否存在程序性之問題,在實證上本研究採用不連續選擇理論中的巢式Logit模式,假設家戶購屋類型選擇決策是有先後順序的過程。在研究方法與設計上,本研究依據模糊語意尺度法(Fuzzy Linguistic scale, FLS)結合巢式Logit模式建立新的理論模式,先就替選方案的指定、解釋變數的定義與效用函數的指定進行說明。接著分別進行家戶購屋類型巢式二項Logit模式(the Nested Binary Fuzzy Model, NBLM)與「在模糊語意下的巢式二項Logit模式(the Nested Binary Fuzzy Linguistic Logit Model, FNBLM)」之參數校估,最後分別對不同模式之校估結果做比較說明。實證結果發現,FNBLM無論在模式的配適能力、預測成功率與概似比統計量檢定等相關統計指標的意義上,皆較NBLM模式具有較佳的整體解釋變異能力。此外,由於包容值(inclusive value)係數值介於0與l之問,顯示本研究所建立之巢式Logit模式為合理選擇決策結構。本研究實證結果亦顯示,在傳統不連續選擇理論中的Logit模式加入模糊語意後,將可以解決傳統住宅消費選擇個體計量經濟模式中具模糊性與不確定性的解釋變數處理的問題,並可作為協助具質化資料與量化分析方法整合的重要參攷。
To improve the IIA problem of multinomial Logit model in discrete choice theory and compare different models as well as test sequential choice process between housing type and unit choice, this paper first suppose that type choice in household purchase is sequential choice process. This paper tries to build a new model by combining Fuzzy Linguistic Scales (FLS) approach with the nested Logit model. In research design, this paper first explains the definition of alternatives, independent variables and utility function. Then this paper processes the comparison of coefficient estimations for type choice in household purchase of the Nested Binary Logit Model (NBLM) and Fuzzy Linguistic Logit Model (FNBLM). The final result is that the goodness-of-fit, success rate of forecasting and likelihood statistic test of model in FNBLM are all better than those NBLM. In addition, due to the inclusive value between 0 and 1, the FNBLM in this paper is the rational structure of choice decision-making. Another finding of this study is that the new model is proposed to be more capable of dealing with the problem of qualitative variables, which is one of the critical issues in quantitative approaches of the housing consumption choice behavior model in household.