Large-scale feature selection using evolved neural networks

摘要: In this paper computational intelligence, referring here to the synergy of neural networks and genetic algorithms, is deployed in order determine a near-optimal network for classification dark formations oil spills look-alikes. Optimality sought framework multi-objective problem, i.e. minimization input features used and, at same time, maximization overall testing accuracy. The proposed method consists two concurrent actions. first identification subset that results highest accuracy on data set feature selection. second parallel process search topology, terms number nodes hidden layer, which able yield optimal with respect selected features. show method, concurrently evolving yields superior compared sequential floating forward selection as well using all together. matrix generalization capacity discovered topology evolved sub-set