| Summary: | 碩士 === 國防管理學院 === 國防資訊研究所 === 94 === By realizing the characteristics and features of things, people can understand, discriminate and classify all things in the world. Therefore, the approach of how to discriminate the difference between things has become more and more important. Although Discrimination Analysis is being applied extensively in the empirical research, most of those did not verify the validation of the hypotheses before data analysis and it could be doubted whether the conclusion with robustness or not. Furthermore, when comparison of discriminate methods, most of researchers always used to adopt specific data, without including general information; however, that way is easy to result in particular conclusions. Hence, for improving this defect the study applies Monte Carlo Method to simulate multivariate normal data and then makes selection of discriminate function can be more generalized.
Keyword: Discriminant Analysis、robustness、Monte Carlo Method
|
|---|
