The aim of this dissertation is to assess the effectiveness of LLMs such as FinBERT and GPT-2 in detecting fraudulent activities in financial reports and statements. This repo provides the code for implementing LLMs, traditional machine learning and deep learning models on the labelled dataset

A repository for the research article titled "DBNex: Deep Belief Network and Explainable AI based Financial Fraud Detection".

Developed and evaluated machine learning and deep learning models for detecting financial fraud.

An Explainable AI (XAI) based machine learning project for financial fraud detection and credit scoring that uses interpretability techniques such as SHAP and LIME to explain model predictions, enhance transparency, and support trustworthy financial decision-making.

identify fraudulent financial transactions using key behavioral patterns in transaction data. It includes comprehensive data preprocessing, model training using scikit-learn, and a Streamlit web app that allows users to input transaction details and receive instant fraud predictions.

Detection of fraudulent credit card transactions via gradient boosting

This project aims to train a Machine Learning Algorithm to recognize financial fraud. In order to do this we use Scikit Learn.

Machine Learning project using PySpark & Streamlit to detect financial fraud

This project detects financial fraud using machine learning techniques. It includes a pipeline from data collection and model training to deployment with a Flask web app. Users can input transaction data and receive real-time fraud predictions instantly. The models are trained in Google Colab and deployed locally using Flask for easy access.

Explainable AI pipeline for financial fraud detection on SEC EDGAR filings using FinBERT, verification, and human-evaluation surveys.

A machine learning project implementing multiple approaches to detect fraudulent credit card transactions.

Privacy-Preserving Edge AI for Financial Fraud Prevention

Production-grade RAG pipeline that translates natural language into executable pandas code for financial fraud analysis over 6M+ PaySim transactions. Multi-model inference (Llama-3-70B + Qwen2.5-Coder-32B) with 4-stage cascading fallback, agentic self-correction, and AST-based security guardrails.