DocumentCode :
3208852
Title :
Chosen problems of designing effective Multiple Classifier Systems
Author :
Wozniak, Michal ; Zmyslony, Marcin
Author_Institution :
Dept. of Syst. & Comput. Networks, Wroclaw Univ. of Technol., Wroclaw, Poland
fYear :
2010
fDate :
8-10 Oct. 2010
Firstpage :
42
Lastpage :
47
Abstract :
Paper presents a brief survey of the main topics connected with Multiple Classifier Systems design task. On the beginning typical topologies are described and then we focus on ensemble and fuser design methods. In the last section we show the limit of different approaches based on weighted voting.
Keywords :
learning (artificial intelligence); pattern classification; pattern recognition; sensor fusion; ensemble design; fuser design; machine learning; multiple classifier system design; pattern recognition; Algorithm design and analysis; Artificial neural networks; Classification algorithms; Optimization; Pattern recognition; Support vector machine classification; Topology; combining classifier; ensemble design; fuser design; multiple classifier systems; pattern recognition;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Computer Information Systems and Industrial Management Applications (CISIM), 2010 International Conference on
Conference_Location :
Krackow
Print_ISBN :
978-1-4244-7817-0
Type :
conf
DOI :
10.1109/CISIM.2010.5643694
Filename :
5643694
Link To Document :
بازگشت