M.A. Thesis:

combination the evolutionary algorithms with rough sets for optimal attribute selection

Nowadays datasets are huge; and for saving the energy we need to less the capacity of information. In datamining there are lots of methods for reducing data dimensionality. We propose a new feature selection strategy based on rough sets and Particle Swarm Optimization (PSO). Rough sets has been used as a feature selection method with much success, that its mathematical basis is discernibility matrix; There is also a technique for removing vagueness and uncertainty in data. but current hill-climbing rough set approaches to feature selection are inadequate at finding optimal reductions as no perfect heuristic can guarantee optimality. On the other hand, complete searches are not feasible for even medium-sized datasets. So, stochastic approaches provide a promising feature selection mechanism. Like Genetic Algorithms, PSO is a new evolutionary computation technique, in which each potential solution is seen as a particle with a certain velocity flying through the problem space. The Particle Swarms find optimal regions of the complex search space through the interaction of individuals in the population. PSO is attractive for feature selection in that particle swarms will discover best feature combinations as they fly within the subset space. In this work particle swarm will discover the best combination of features flying in the feature space, after evaluating by cost function named rough set. Experimentation is carried out, using UCI data, which compares the proposed algorithm with a GA-based approach and other deterministic rough set reduction algorithms. The results show that PSO is efficient for rough set-based feature selection.

Keywords: Feature selection; Rough Sets; Reduct; Genetic Algorithms; Particle Swarm Optimization