This project is an end-to-end machine learning solution for predicting blueberry yield based on various environmental and biological factors. Using Python and Flask for the back-end and Bootstrap for the front-end, it incorporates data ingestion, transformation, model training, and prediction stages. The prediction model is powered by CatBoost Algo

AI-powered e-waste valuation system that predicts device resale value, estimates recoverable materials, and recommends whether to resell or recycle electronic devices.

A Crop Yield Prediction regression model built using CatBoost, enhanced with SHAP-based interpretability and optimized through hyperparameter tuning with Optuna.

This project focuses on forecasting cryptocurrency prices to aid in investment decisions, risk management, and market understanding. It aims to enhance predictive models for digital asset markets, providing more reliable insights for investors and traders.

Technical audit of Automated Decision System for Fairness and Bias

Interconnect seeks to forecast customer churn by analyzing package choices and contracts. If a customer plans to leave, they're offered unique codes and special packages to foster loyalty.

The project predicts the probability of loan default using various financial features of customer. I applied SMOTENN by combining SMOTE and Edited Nearest Neighbor (ENN) to handle class imbalance. Logistic Regression, Random Forest and CATBOOST models have been apllied and evaluated based on accuray, F1 score, ROC-AUC score.