Lazy learner and PCA: An evolutionary approach

In our work, we acquaint a novel methodology to classify information in different datasets. Our strategy depends on K-nearest neighbor (KNN) classification and utilizes Principal Component Analysis (PCA) as information decrease approach, conquering the constraints of conventional KNN characterization on huge information. PCA is utilized to perform information dimensionality decrease while the KNN lazy learner is utilized to perform classification. The execution of the KNN technique, as far as accuracy and classification time, is resolved as a component of the pressure rate accomplished in the PCA preprocessing stage. KNN classifier gives a tremendous change in this specific case, subsequent to as no preparation is required, new items can be included at whatever time. However, KNN lazy learner has high calculation cost depending on number of instances. In order to make KNN lazy learner work possible for any size of dataset, it is important to preprocess the dataset to decrease the number of instances. Here, PCA is applied to reduce the number of transactions which decreases the computation cost and execution time additionally accomplishes close exactness result than without PCA.