In the real-world domain, many learning models faces challenge in handling the imbalanced classification problem. Imbalanced classification is a scenario where the number of data points in minority class is much lower than that of the majority. Our primary concern is the minority class, which is often neglected by learning models while predicting the values. This problem can be tackled at the data-level by using resampling techniques. In this research, hybrid of Synthetic Minority Oversampling Technique (SMOTE) and Neighborhood Cleaning Rule (NCL) is proposed to balance the data points of the classes. For experiment real-world dataset of credit card transaction has been utilized where the fraudulent (or malefactor) transaction needs to be identified. This imbalanced dataset after resampling is classified by using the logistic regression model. The experimental results depict that the learning model has correctly identified the malefactor in the balanced dataset than the original dataset. Through balancing the datasets, the proposed technique aims to enhance the performance of the learning model in order to correctly identify the cases of the minority class.