Improving Classifier Efficiency by Expanding Number of Functions in the Dataset

Abstract

An order task starts with a data set where the class tasks are known. A dataset with fewer elements, which causes the aggregation run of a classifier to decrease. This paper suggests two quality development strategies for a data set. The class plausibility construction technique is used for highlights with a weak crossing region and the manufacturing component development strategy is used for elements with a high crossing region. An attempt is made to analyses the presentation of the proposed technique using four data sets with two classes and four data sets with several classes with different stroke sizes. The results show that the proposed technique has a higher order execution with Support Vector Machine (SVM) classifier when compared with K-nearest neighbor (KNN) classifier.