Artificial Bee Colony Approach for Approach for Approach for Approach for Optimizing Feature Selection
Bee Colony, Artificial Bee Colony, Shunmugapriya Palanisamy, Selvadurai Kanmani · 2012
High dimensionality of the feature space affects the classification accuracies and the computational complexity due to redundant, irrelevant and noisy features present in the dataset. Feature Selection extracts the relevant and most useful information and helps to speed up the task of classification. Feature selection is seen as an optimization problem because selecting the appropriate, optimal feature subset is very important. The Artificial Bee Colony algorithm is a famous meta-heuristic search algorithm used in solving combinatorial optimization problems. This paper proposes a new method of feature selection, which uses the ABC algorithm to optimize the selection of features. Ten UCI datasets have been used for evaluating the proposed algorithm. Experimental results show that, ABC-Feature Selection has resulted in optimal feature subset configuration and increased classification accuracies up to 12% compared to the classifier and standard ensembles.