Neste projeto, estão sendo utilizados os algoritmos de Machine Learning, Decision Tree e KNN, para detecção de fraudes em cartão de crédito. 💳

Système de détection de fraude bancaire par IA - Précision 99.96%

Este projeto tem como objetivo desenvolver e avaliar modelos de machine learning para detecção de fraude em transações financeiras online, utilizando o dataset público IEEE-CIS Fraud Detection (Kaggle).

Pipeline complet de détection de fraude bancaire utilisant XGBoost (F1: 85%). Inclut le nettoyage automatisé, l'équilibrage des données (SMOTE) et l'optimisation des seuils de décision pour minimiser les faux positifs

End-to-end AML detection pipeline on the Elliptic Bitcoin transaction graph, comparing tabular baselines (cuML/sklearn) and GNNs (PyTorch Geometric) with reproducible phases, metrics, and curated results.

Proyecto Detección Fraude

An end-to-end fraud investigation demo built with XGBoost, SHAP, RAG, and Gemini API. A Streamlit app accepts a transaction, scores it with a trained classifier, explains the prediction with SHAP, retrieves relevant fraud typology from a local knowledge base, and produces a structured investigator report via a Gemini-powered agent.