Using a hybrid model of Clustering Analysis, Rough Set Theory and Support Vector Machines to predict listing corporate failure
| Autor: | Hsiao-Jung Liu, 劉曉蓉 |
|---|---|
| Rok vydání: | 2009 |
| Druh dokumentu: | 學位論文 ; thesis |
| Popis: | 97 The high social costs associated with corporate failure have spurred searches for better theoretical understanding and prediction capability. The accuracy of classification became an important issue because it is the core of prediction. In this paper, we investigate a hybrid approach to corporate failure prediction, using K-mean clustering, Rough Set Theory, and Support Vector Machines to construct a corporate failure prediction model. The study used data from Taiwan Economic Journal (TEJ) database for the period 1971 to 2009. The hybrid model we developed in this research is 89% accurate on a validation sample as compared to the original rough sets model which was 78% accurate. The K-mean clustering and Support Vector Machines help improving the accuracy of Rough Set Theory. These findings indicate that K-mean clustering and Support Vector Machines coupled with Rough Set Theory can be an efficient and effective hybrid modeling approach. |
| Databáze: | Networked Digital Library of Theses & Dissertations |
| Externí odkaz: |
|