Fraud Detection Analysis using Machine Learning

Project Overview

This project analyzes financial transaction data to identify fraudulent transactions using Machine Learning techniques. The workflow includes data cleaning, exploratory data analysis (EDA), feature engineering, model training, evaluation, and feature importance analysis. A Random Forest Classifier was used to classify fraudulent and non-fraudulent transactions based on transaction patterns and account behavior.

Technologies Used

• Python
• Pandas
• NumPy
• Matplotlib
• Seaborn
• Scikit-Learn
• Machine Learning

Key Results

• Built a fraud detection model using Random Forest Classifier.
• Performed data cleaning, feature engineering, and exploratory data analysis.
• Identified important fraud indicators such as balance differences, transaction amount, and account balances.
• Evaluated model performance using classification metrics.

Project Workflow

• Data Loading and Preprocessing
• Missing Value Analysis
• Exploratory Data Analysis (EDA)
• Outlier Detection
• Feature Engineering
• Correlation Analysis
• Train-Test Split
• Model Development
• Model Evaluation
• Feature Importance Analysis

Key Analysis Performed

• Fraud vs Non-Fraud Distribution Analysis
• Transaction Type Analysis
• Transaction Amount Distribution
• Correlation Heatmap Analysis
• Feature Importance Evaluation

Machine Learning

The dataset was prepared using feature engineering and train-test splitting before training a classification model to identify fraudulent transactions.

Results

The model was evaluated using standard machine learning evaluation techniques to assess its ability to detect fraudulent transactions accurately.

Dataset Note

The original dataset was not uploaded due to its large size. The analysis was performed on a sampled subset for efficient processing and model development.

Skills Demonstrated

Python, Data Cleaning, Exploratory Data Analysis, Data Visualization, Feature Engineering, Machine Learning, Model Evaluation, Fraud Detection.