台灣花卉價格分佈的不均等衡量值分析

Abstract

本論文比較四種分析台灣花卉價格分佈的不均等衡量值：羅侖茲曲線、吉尼係數、泰爾指數及申氏不均等指標。其中羅侖茲曲線用圖形來呈現花卉價格分佈的整體特色。相對而言，吉尼係數、泰爾指數及申氏不均等指標則是用數字來表達。從資料倉儲角度來看，此比較過程中泰爾指數明顯地展現出特優於其它的三種分析方法。換言之，台灣的花卉價格資料取自具有維度模式的資料倉儲。這取得資料的過程常需在維度模式裡進行上捲與下挖的操作。泰爾指數的可分解性特色使得不均等衡量值能在不同的層級中保持一致性。根據此項數學性質，本研究亦實作出具有泰爾指數特色的一套地理資訊系統。

Four inequality measurements for analyzing the Flower Price Distribution (FPD) in Taiwan have been compared: Lorenz curve, Gini coefficient, Theil index, and Sen’s measure of inequality. The Lorenz curve method offers a graphic representation for showing a holistic feature of FPD. In contrast, Gini coefficient, Theil index, and Sen’s measure of inequality methods are numerical representations. In this comparison, the Theil index approach obviously shows an inherent advantage over the other three methods from a data warehouse viewpoint. In other words, the data of flower prices in Taiwan are fetched from a data warehouse based on dimensional model. The data retrieving process needs to drill down or roll up in the dimensional model. The Theil index with a unique decomposability property shows a consistent way for representing the inequalities among different levels. According to this mathematical property, a geographic information system with Theil index feature has been implemented.