EASY ENSEMMBLE WITH RANDOM FOREST TO HANDLE IMBALANCED DATA IN CLASSIFICATION

Abstract

Imbalanced data might cause some issues in problem definition level, algorithm level, and data level. Some of the methods have been developed to overcome this issue, one of state-of-the-art method is Easy Ensemble. Easy Ensemble was claimed can improve model performance to classify minority class, and overcome the deficiency of random under- sampling. In this paper we discussed the implementation of Easy Ensemble with Random Forest Classifiers to handle imbalance problem in credit scoring case. This combination method is implemented in two datasets which taken from data science competition website, finhacks.id and kaggle.com with class proportion within majority and minority is 70:30 and 94:6. The results showed that resampling with Easy Ensemble can improve Random Forest classifier performance upon minority class. Recall on minority class increased significantly after the resampling. Before resampling, the recall on minority class for the first dataset (finhacks.id) was 0.49, and increased to 0.82 after the resampling. Similar results were obtained for the second data set (kaggle.com), where the recall for the minority class was increased from just 0.14 to 0.73.