🚀 Explore Gradient Boosting techniques and low-default modeling for financial data science, enhancing strategies for tackling extreme class imbalance.

🤖 Build an AI-driven loan eligibility and risk scoring system to facilitate smarter loan decisions with advanced machine learning techniques.

🏦🤖FinChurn is an advanced Financial Machine Learning system designed to predict customer churn and detect fraudulent activities with high accuracy. It includes a complete end-to-end ML workflow covering data preprocessing, exploratory analysis, class imbalance handling (SMOTE)

Recursive Partitioning for Structural Equation Models

GraphSAGE + XGBoost ensemble for DNS malware detection. Achieved 97.43% accuracy on 200K-domain Zenodo dataset. Internship project.

production-grade Telecom Customer Churn Prediction System (ML • SMOTE • Streamlit • AWS EC2 • CI/CD)

Predictive Customer Churn Analysis and Strategic Segmentation using LightGBM and K-Means. Features an interactive Streamlit dashboard for actionable retention strategies.

Prediction of pathologies based on biomechanical data using machine learning models (logistic regression, random forest)

YouTube video category classification using TF-IDF, PhoBERT, E5, BGE, and SVM/Random Forest. (Phân loại thể loại video YouTube từ mô tả bằng TF-IDF, PhoBERT, E5, BGE.)

A machine learning project using Random Forest and XGBoost to predict user churn and enable proactive customer retention strategies.

FraudWatch is a machine learning-based credit card fraud detection system that uses a Random Forest classifier. It visualizes model performance with an interactive confusion matrix heatmap. The system is deployed as a user-friendly Flask web application. 📊

End-to-end Telecom Customer Churn Prediction Pipeline (Python, MySQL, ML, Power BI)

A sentiment analysis project on Twitter tweets using Python for text preprocessing and visualization. The project includes Exploratory Data Analysis (EDA) and sentiment classification using Logistic Regression and Random Forest models.

This project predicts whether an Airbnb listing will achieve a perfect (100%) rating score using data on listing features, host behavior, and property characteristics.

It is important that credit card companies are able to recognize fraudulent credit card transactions so that customers are not charged for items that they did not purchase.

An interactive machine learning app that predicts UFC fight outcomes using real fighter data (1996–2024). Built with Python, Scikit-learn, and Gradio — this project models fight winners, methods, and statistics with real-world tested accuracy.

Baseline machine-learning models for classifying human activity using wearable sensor features. Includes preprocessing, feature selection, evaluation metrics, and visualization

A machine learning project for diabetes prediction using Logistic Regression, SVM, and Random Forest. After model tuning, Random Forest achieved the best performance with 78.8% accuracy and 86.0% ROC-AUC, improving early diabetes detection.

Genetic Variant Classification using machine learning

This repository contains all the basic code related to machine learning and deep learning...