Skip to content

Krish5986/Financial-Fraud-Detection

Folders and files

NameName
Last commit message
Last commit date

Latest commit

Β 

History

3 Commits
Β 
Β 
Β 
Β 

Repository files navigation

Financial Fraud Detection Pipeline

πŸ“Œ Project Overview

This project involves building an end-to-end Machine Learning pipeline to detect fraudulent transactions in a massive financial dataset. The goal was to develop a robust model capable of identifying fraud patterns within 6.3 million transaction records, focusing on precision and minimizing false positives.

This analysis was conducted as part of a Data Analyst case study for Accredian.

πŸš€ Key Features

  • Large-Scale Data Processing: Efficiently handled and cleaned a dataset of 6.3 million rows involving financial transfer logs.
  • Advanced Feature Engineering:
    • Created new features to track balance discrepancies (origin vs. destination accounts).
    • Calculated Variance Inflation Factor (VIF) to detect and remove multicollinearity between independent variables.
  • Model Architecture: Implemented a Random Forest Classifier optimized for imbalanced data.
  • Strategic Insights: Derived actionable insights on fraud prevention and real-time monitoring strategies.

πŸ› οΈ Tech Stack

  • Language: Python
  • Libraries: Pandas, NumPy, Scikit-learn, Matplotlib, Seaborn
  • Techniques: SMOTE (imbalance handling), VIF Analysis, Correlation Matrix

πŸ“Š Results

  • Precision: Achieved 95% precision on the fraud class, ensuring reliable detection with minimal false alarms.
  • Key Drivers: Identified that specific transaction types (CASH_OUT and TRANSFER) were the primary vectors for fraud.

Created by Krish

About

No description, website, or topics provided.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors