- 作者: 徐悌
- 作者服務機構: 大業工學院事業經營研究所
- 中文摘要: 本研究測試二個不同的線型程式方法包括MMD及MSD模型以決定他們在二個團體區分問題上會員分組的有效性。三組由小到100個到大到2600個不同現實生活中的資料來做上述的測試。用這二種方式測試出來的結果與用傳統統計方式:像區別分析及Logistic的迴歸分析得出來的結果相比較。比較的結果發現在二個團體會員分組的問題上,線型程式方法還是無法取代傳統的統計方法。
- 英文摘要: Two different LP approaches, including MMD model and MSD model, were tested for theireffectiveness in classifying memberships into two-group situations. There were three different real-world datasets used in the test ranging from as small as 100 to as large as 2600 subjects. The results were then comparedto the results obtained using the same data sets from the conventional statistical techniques, such asdiscriminant analysis and logistic regression analysis. Findings could not be used to substantiate its support for those who advocated the use of LP techniquesas substitutes for the conventional statistical methods.
- 中文關鍵字: Linear programming; parametric methods;discriminant analysis; logistic regression; group membership classification
- 英文關鍵字: --