fraud detection

This project focuses on building a Fraud Detection System using data analysis and machine learning techniques. The code implements a complete pipeline starting from data preprocessing and feature engineering to model training, evaluation, and optimization.

First, feature engineering is applied by extracting the transaction hour from the timestamp and calculating the time gap between the transaction and the last login. These features help capture user behavior patterns that may indicate fraudulent activity.

The dataset is then prepared by encoding categorical variables using LabelEncoder and splitting the data into training and testing sets. A classification model is built using DecisionTreeClassifier to predict whether a transaction is fraudulent or not.

To address class imbalance, the project applies SMOTE, which improves the model’s ability to detect fraud cases. The model is evaluated using key performance metrics such as accuracy, precision, recall, F1-score, and confusion matrix.

Additionally, hyperparameter tuning is performed باستخدام GridSearchCV to optimize the model’s performance by selecting the best combination of parameters.

Finally, the model is used for inference on unseen data to predict whether a transaction is Fraud or Not Fraud, demonstrating its practical applicability in real-world scenarios.