Identifying the Contributors of the Multivariate Variability Control Chart using Hierarchical Support Vector Machines
碩士 === 輔仁大學 === 應用統計學研究所 === 99 === How to identify the contributors is a crucial problem in multivariate statistical control. A rarely recognized property of this problem is that the number of possible classes can be very large. In this case, a classifier with hierarchical structure might be an alt...
| Main Authors: | , |
|---|---|
| Other Authors: | |
| Format: | Others |
| Language: | zh-TW |
| Published: |
2011
|
| Online Access: | http://ndltd.ncl.edu.tw/handle/73449833342083581834 |
| Summary: | 碩士 === 輔仁大學 === 應用統計學研究所 === 99 === How to identify the contributors is a crucial problem in multivariate statistical control. A rarely recognized property of this problem is that the number of possible classes can be very large. In this case, a classifier with hierarchical structure might be an alternative that can achieve better performance. In this study, a hierarchical classifier called HSVM is used for investigating this idea. Besides, the way of tuning the associated parameter is also discussed in this research. The experiment results show that HSVM has the best performance by comparing with other popular powerful classifiers including regular SVM.
|
|---|