Designing a Model to Handle Imbalance Data Classification Using SMOTE and Optimized Classifier

Abstract

Classification is a supervised learning process that is used to predict target class for an input variable. In an imbalanced dataset, the total count of examples of minority class is comparatively less than the total count of majority (negative) class. Various problems accompany during classification of imbalanced data. To eliminate this problem, there is a need to balance the dataset as well to achieve higher accuracy by optimizing the parameters of a classification model. The proposed system attempts to sample the imbalance dataset into a balanced dataset by using SLS (Safe-Level Synthetic Minority Oversampling Technique) and ADASYN (Adaptive Synthetic Sampling Approach) SMOTE techniques. On this balanced data, for classification, KNN (K-Nearest Neighbors) and SVM (Support Vector Machine) classifiers are used. The system proposes to improve the performance of the classifiers using metaheuristic Cuckoo Search optimization techniques for determining hyperparameters.