Subscribe Now Subscribe Today
Science Alert
 
FOLLOW US:     Facebook     Twitter
Blue
   
Curve Top
Information Technology Journal
  Year: 2013 | Volume: 12 | Issue: 23 | Page No.: 7754-7757
DOI: 10.3923/itj.2013.7754.7757
A New Fuzzy Combination Method Based on Parametric Triangle Norm
Jia Pengtao and Liang Shuhui

Abstract:
Triangular norms can improve the generalization capability of pattern classification problems, so they are introduced to improve the performance of ensemble learning. This study pays attention to the different impacts of base classifiers and put forward a new fuzzy combination method for ensemble learning. It is based on parametric triangle norms. Firstly, a new combination model was constructed. The base learners in it are set with different weights. Secondly, a set of fuzzy combination rules were generated according to the parametric triangle norms. Thirdly, genetic algorithm was used as the parameters estimation module of the new fuzzy rules. Finally, experiments were conducted on seven different datasets from the University of California, Irvine machine learning repository (UCI). The experimental results show that the fuzzy rules generated by the new combination method have better performance than the base learners and the fixed combination rules. And when set proper weights to base learners, the new fuzzy rules obtain better performance than the ones set a same weight to base learners.
PDF References Citation Report Citation
How to cite this article:

Jia Pengtao and Liang Shuhui, 2013. A New Fuzzy Combination Method Based on Parametric Triangle Norm. Information Technology Journal, 12: 7754-7757.

DOI: 10.3923/itj.2013.7754.7757

URL: https://scialert.net/abstract/?doi=itj.2013.7754.7757

 
COMMENT ON THIS PAPER
 
 
 

 

 
 
 
 
 
 
 
 
 

 
 
 
 
 

       

       

Curve Bottom