Handling unbalanced classes using assembly methods

The present work concerns decision optimization using ensemblistic methods for processing unbalanced databases. To achieve this, we used ensemblistic methods, which are based on the homogeneous combination of predictions or classifiers for better generalization.In our project, we used the Credit Card Fraud Detection database to generate and evaluate the proposed model. We also chose the random forest combination method, which combines several decision trees and applies the majority voting strategy to generate an optimal prediction.The aim of our study is to build a prediction model using assembly methods to improve the performance of an individual classifier in dealing with unbalanced datasets.To achieve our goal, apart from the random forest combination method used, we also used the subsampling and oversampling methods to achieve the same results and finally draw a conclusion on the three methods used and this we implemented with the phyton programming language.