Credit_card_fraud_detection

This Credit Card Fraud Detection project involves using the Logistic Regression algorithm on a set of features to detect whether a credit card transaction is fraudulent or not. Raw data is then normalized and an attempt at solving the problem of a manageable number of samples and instances per class to enhance the training process by applying SMOTE. The model was trained and evaluated, achieving the following metrics:
Precision: 0. 97
Recall: 0. 92
F1-Score: 0. 95

Looking at the confusion matrix and the classification report we can see that the algorithms have performed quite well. Additional analysis that encompasses confusion matrix and ROC curve offer understanding of the classifier’s performance at discriminating fraud.