KNN Model Selection using Modified Cuckoo Search Algorithm

In this paper, an automated model selection approach guided by Cuckoo search is proposed for k-nearest neighbor (KNN) learning algorithm. The performance of KNN mostly depends on the value of k and the distance metric used. The values of these parameters are computed by optimizing an objective function designed for measuring the classification accuracy of KNN. Cuckoo search being an efficient optimization technique has been used to optimize the value of k and select a distance metric among Euclidean, city-block, cosine, and correlation metrics for KNN classifier. Numerous experiments have been conducted on benchmark datasets to validate the performance of the proposed approach. Results demonstrate that the approach is practical and very efficient.