Abstract | ||
---|---|---|
Weighted combination model with appropriate weight vector is very effective in multiple classifier systems. We presented a method for determining the weight vector by particle swarm optimization in our previous work, which called PSO-WCM. A weighted combination model, PSO-LS-WCM, was proposed in this paper to improve the classification performance further, which obtained the weighted vector by particle swarm optimization with local search. We describe the algorithm of PSO-LS-WCM in detail. Seven real-world problems from UCI Machine Learning Repository were used in experiments to justify the validity of the approach. It was shown that PSO-LS-WCM is better than PSO-WCM and the other six combination methods in literature. |
Year | DOI | Venue |
---|---|---|
2011 | 10.1007/978-3-642-21524-7_29 | ICSI |
Keywords | Field | DocType |
combining classifier,particle swarm optimization,weighted vector,weighted combination model,multiple classifier system,uci machine learning repository,weight vector,combination method,appropriate weight vector,local search,classification performance | Particle swarm optimization,Mathematical optimization,Pattern recognition,Computer science,Weight,Multi-swarm optimization,Artificial intelligence,Local search (optimization),Classifier (linguistics),Particle,Machine learning,Metaheuristic | Conference |
Volume | ISSN | Citations |
6729 | 0302-9743 | 0 |
PageRank | References | Authors |
0.34 | 5 | 1 |
Name | Order | Citations | PageRank |
---|---|---|---|
Liying Yang | 1 | 11 | 7.05 |