Mushroom Classification – Kaggle Competition

Developed a high-performance machine learning model to classify mushrooms as edible or poisonous using a large-scale dataset (3M+ rows). Performed end-to-end data preprocessing, exploration, and visualization, with One-Hot Encoding for categorical features. Built and optimized an XGBoost model using GridSearch, achieving ~99% accuracy on training and test data and ~97% on Kaggle, demonstrating strong generalization and reliability in a competitive environment.