This paper presents a review of some of the most recent evolutionary algorithms used for solving feature selection based upon previously published research on feature selection. In addition, we discuss various research issues relating to each of the presented evolutionary algorithm. Evolutionary algorithms present several advantages over traditional search such as they require less domain-specific information. Such advantages have made them very popular within feature selection as explained in this paper. This paper covers the first part only of the evolutionary algorithms for the feature selection problem due to the limitation of the number of pages. The references cited in this paper cover the major theoretical issues, and provide access to the main branches of the literature dealing with such methods.

Abstract-Many applications such as pattern recognition require selecting a subset of the input features in order to represent the whole set of features. The aim of feature selection is to remove irrelevant or redundant features while keeping the most informative ones. In this paper, an ant colony system approach for solving feature selection for classification is presented. The proposed algorithm was tested using artificial and real-world datasets. The results are promising in terms of the accuracy of the classifier and the number of selected features in all the used datasets. The results of the proposed algorithm have been compared with other results available in the literature and found to be favorable.

High accuracy and short amount of time are required for the solutions of many classification problems such as real-world classification problems. Due to the practical importance of many classification problems (such as crime detection), many algorithms have been developed to tackle them. For years, metaheuristics (MHs) have been successfully used for solving classification problems. Recently, hybrid metaheuristics have been successfully used for many real-world optimization problems such as flight scheduling and load balancing in telecommunication networks. This chapter investigates the use of this new interdisciplinary field for classification problems. Moreover, it demonstrates the forms of metaheuristics hybridization as well as designing a new hybrid metaheuristic.

Due to the need for the dimensionality reduction in many applications such as real-world and large scale applications, this paper demonstrates it, and covers it from all of its perspectives such as its components, its importance, and where it is needed. Moreover, this paper demonstrates the impact of irrelevant features on the performance of many predictors such as decision tree, Naï ve Bayes, nearest neighbor, and support vector machines. This is because the performance of these predictors degenerates when provided with data containing irrelevant features.

No abstract