Fuzzy Artificial Fish Ensemble Extreme Learning Machine

摘要

Neural networks have been largely applied into many real world pattern classification problems, and the combination of neural networks to overcome the natural limitations of single classifiers is one strategy to achieve better accuracy. Due to the need of having accurate classifiers with different knowledge for the same problem, ensembles need to measure their diversity. In this work we employ a Fuzzy Adaptive Modified Artificial Fish Swarm Algorithm for evolving Extreme Learning Machine Ensembles. The ensemble selection is performed by the use of a hierarchical clustering algorithm, and two diversity measures. Experimental results show that the proposed method achieved promising results.