Popular GitHub Repositories for Fraudulent Transactions

This notebook tries to make fraud/not fraud predictions on a transactions dataset with highly imbalanced data.

? Creates a new gym environment for credit-card anomaly detection using Deep Q-Networks (DQN) and leverages Open AI's Gym toolkit to allocate appropriate awards to the RL agent.

The objective of my experiment is to analyze the performance of Random Forest, Naive Bayes, Logistic Regression, and K-nearest neighbors machine learning models for evaluation of the utility of synthesized data for fraud detection.

Geração de dados sintéticos para transações fraudulentas de e-commerce utilizando o framework LangChain e modelos da OpenAI. Esse material foi elaborado durante a disciplina Processamento de Linguagem Natural na UFABC.