Behavioral Analytics for Detecting Anomalies in Financial Transactions problem Statement: Implement and improve behavioral analytics solutions that can detect anomalies in financial transactions, providing enhanced security against fraudulent activities.
Working Principle: Our chosen anomaly detection approach employs an isolated forest machine learning model, leveraging unsupervised ensemble learning. This model isolates anomalies within finance transaction data subsets, utilizing feature influence and importance, and supports hyperparameter tuning for optimal performance.
Dataset : Attached with the Repo
Approach: The overview of the approach focuses on identifying anomalies and it’s reasoning patterns (cause-effect) by following a series of steps on achieving data collection, integration, training & tuning the model, feature engineering, behavioral profiling and authentication reaction set-up
Libraries used: sklearn.ensemble - isolationForest sklearn.model_selection - GridSearchCV tkinter pandas
For further information u can refer presentation link : https://www.canva.com/design/DAF4_M5qGVg/jFzV3C8dUwcYQ20C5NfuDA/edit?utm_content=DAF4_M5qGVg&utm_campaign=designshare&utm_medium=link2&utm_source=sharebutton
and MFA