AoedeUse these settings →
This ML Toolkit portfolio demonstrates the development of practical machine learning systems with measurable, real-world impact. The first project focuses on Diabetes Prediction, where three classification algorithms — K-Nearest Neighbors, Decision Tree, and Random Forest — were systematically evaluated. Random Forest emerged as the top-performing model, achieving 84% accuracy alongside the lowest RMSE, validated through rigorous metrics including Accuracy, F1 Score, and RMSE. The workflow encompassed data preprocessing, exploratory data analysis, and feature engineering, all built on Scikit-learn, Pandas, and NumPy. The second project is a Facial Emotion Recognition system powered by a custom Convolutional Neural Network trained on the FER-2013 dataset. Leveraging TensorFlow and Keras, the model delivers real-time emotion predictions through an interactive dashboard deployed via Streamlit — bringing deep learning capabilities directly into a production-ready user interface.