Purpose

Implement machine learning models which are able to recognize fraudulent credit card transactions so that customers are not charged for items that they did not purchase.

Content

The dataset contains transactions made by credit cards in September 2013 by European cardholders (Download here in Kaggle: https://www.kaggle.com/datasets/mlg-ulb/creditcardfraud).

This dataset presents transactions that occurred in two days, where have 492 frauds out of 284,807 transactions. The dataset is highly imbalanced, the positive class (frauds) account for 0.172% of all transactions. It contains only numerical input variables which are the result of a PCA transformation due to confidentiality.

Solution

• Preprocess data (drop missing values, normalize data, ...)
• Apply classification algorithms into original imbalanced dataset to evaluate their performance.
• Apply undersampling, oversampling method to solve imbalanced issue.
• Re-apply classifications algorithms into undersampled/oversampled dataset to see if the performance of models is improved.
• Trying apply ensemble methods (Bagging, RandomForest, Boosting, ...) into undersampled/oversampled dataset to evaluate performance.