Feature selection using cuckoo search algorithm for object classification
In any object classification system, there is a need to extract features and use them to classify the objects, most of the extracted features have succeed to classify some object but failed to classify others. Feature selection is a general problem used for dimensionality reduction purposes. Feature selection aims to select important features. The major objective of the this paper is to use the Binary Cuckoo Search(BCS) to select important features from the set of the extracted features. The feature extracted were grey level features ( texture features ), the size features, shape features and the optimal feature has been selected by BCS. Support vector machine (SVM) classifier used in BCS as a fitness function. The cuckoo search algorithm select only seven feature from (25) features where performed average accuracy 92% and improve the classification time from ( 1.55 ) second at average to ( 57 ) millisecond at average.