Hybrid Approach for Imbalanced Classification with Deep Neural Network

Work on machine learning, especially deep learning, really depends on the quality of data. Data imbalance is one of the problems that we face when applying machine learning to real-world problems. With research in this area is drawing more attention from academics and even industry. The class imbalance problem raises difficulties because of the skewness of the data distribution, since machine learning depends on the data generalization. To address this imbalanced data problem, we adopt a hybrid (algorithm and data) approach that consists of data manipulation and weighted loss function in this thesis. we propose Ripple-SMOTE as a novel oversampling method to generate synthetic data for preprocessing. A deep neural network and the weighted loss function is applied so it will not treat all classes equally. This research uses a pre-trained model and fine tune it to improve the classification accuracy. In this thesis, we report the evaluation results using imbalanced data sets based on MNIST, CUReT texture set, and Malware data set, and show that our approach significantly improves the performance in imbalanced data cases and outperforms the conventional approaches, especially in handling minority classes [1]