Enhancing Performance of Lazy Learner by Means of Binary Particle Swarm Optimization

Machine Learning is the area of Artificial Intelligence that deals with develop techniques capable of learning, that is, automatically extract knowledge underlying information. Together with statistics, it constitutes the heart of the intelligent analysis of the data. The principles followed in machine learning and in the mining of data are the same: the machine generates a model from examples and use it to solve the problem. This paper elaborates the hybrid technique K nearest neighbor-Principal Component Analysis (KNN-PCA) for classification on different data base by UCI, further this paper proposes a Binary Particle Swarm Optimization (BPSO) based attribute selection algorithm with objective of higher accuracy and lower training overhead. The experimental set up has been developed with UCI dataset using KNN lazy learner. The accuracy, precision, recall and F-score has been calculated as an evaluation parameter and result compared with BPSO optimized structure, the proposed structure claims higher accuracy the traditional approach.