Research to determine the fuzzy measure system of multiple classifiers based on fuzzy integral fusion

Fuzzy integral is an aggregation tool for classification, which is used to improve the accuracy and robustness of the fusion of multiple systems. Multi-classifier fusion, based on Fuzzy Integrals measure system, will have a great impact on the performance of the fusion system. If well defined, the f...

Full description

Saved in:
Bibliographic Details
Main Authors: Yong-Hua Cai, Bo Wu, Bao-Zhu Chen, Tie-Song Li
Format: Conference Proceeding
Language:English
Subjects:
Abstracts
Accuracy
Classification algorithms
Fuzzy integral
Fuzzy measure
Genetics
Multiple classifier fusion
Time measurement
Online Access:Request full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Fuzzy integral is an aggregation tool for classification, which is used to improve the accuracy and robustness of the fusion of multiple systems. Multi-classifier fusion, based on Fuzzy Integrals measure system, will have a great impact on the performance of the fusion system. If well defined, the fuzzy measures could markedly improve the classification accuracy; conversely, it may even result in less accuracy than a single classifier. Given the fusion of classifier, this paper firstly analyzes the impact of fuzzy measure on the classification result, and points out the multiple classifiers fusion system has the ability of a certain error correction. For example, even though all classifiers are wrong, fuzzy integral fusion system is still possible to classify the sample correctly.
ISSN:2160-133X
DOI:10.1109/ICMLC.2014.7009706