DATA DIMENSIONALITY REDUCTION BY GENETIC ALGORITHMS

Pattern recognition generally requires that objects be described in terms of a set of measurable features. Multi-classification process can be significantly enhanced byselecting an optimal set of the features used as input for the training operation. The selectionand quality of the features representing each pattern have aconsiderable bearing on the success of subsequent pattern classification .The selection of such a subset willreduce the dimensionality of the data samples and eliminate the redundancy and ambiguity introduced by some attributes. Here, we present some approaches to both feature selection and feature extraction using some genetic algorithms.