Credit Card fraud detection is a dataset in Kaggle. Here is my solution for this problem. I have used both undersampling and oversampling techniques to find the best solution. The resampled data has been tested on 6 models namely: Logistic Regrression, K Neighbor Classifier, Decision Treee classifier, Gradient Boosting Classifier, Random Forest Classifier and Support Vector Classifier. Tomek link and SMOTE is used for undersampling and oversasmpling respectively. Conclusion: On both the forms of resampled data Random Forest Classifier performed the best for this dataset.