Hybrid decision tree architecture utilizing local SVMs for multi-label classification
HAIS'12 Proceedings of the 7th international conference on Hybrid Artificial Intelligent Systems - Volume Part II
| Hi-index | 0.00 |
Data mining has many topics such as classification, clustering, association, prediction, etc. Recently, classification problem is the research hotspot and decision tree is one of the most widely used classification methods, where C4.5 is one favorite algorithm. According to the disadvantages of conventional support vector machine (SVM), a SVM based decision tree (SVMDT) is introduced and modified by using equivalent distance as the class separability measure and includingthe consideration of "local class cluster" problem. At last the modified SVMDT is used to make diagnosis analysis of an experimental rotor kit.