A study of feature selection approaches for classification

Abstract

Classification approaches need to be improved in the domain of medical applications. Feature selection for classification is a very active research field in data mining Selecting only relevant features before classification reduces the workload of classifier using feature selection. Feature selection optimizes the classification performance. It finds only useful features and removes non- relevant features, and reduces the input dimensionality which simplifies the implementation of the classifier and speed up the processing rate. Researchers have introduced many feature selection algorithms with different selection criteria. However, it has been found that no single criterion is best for all applications. A hybrid approach for feature selection called based on genetic algorithms (GAs) that employs a target learning algorithm to evaluate features. The merits of this method includes t multiple feature selection criteria and find small subsets of features that perform well for the target algorithm.