A collection of machine learning projects implemented with Python, NumPy, Pandas, and Matplotlib. The projects cover regression, classification, dimensionality reduction, and ensemble learning.
- Implemented linear and generalized linear regression models from scratch using NumPy and Pandas to analyze movie ratings and housing prices. Applied gradient descent, stochastic gradient descent, normal equations, feature transformations, mean squared error evaluation, and model visualization.
- Built binary and multiclass logistic regression models from scratch using NumPy and Pandas to classify student success and iris species. Applied gradient descent, feature scaling, stratified sampling, 5-fold cross-validation, and decision-boundary visualization.
- Implemented PCA and Linear Discriminant Analysis from scratch using NumPy and Pandas for image representation and classification. Applied image vectorization, data centering, dimensionality reduction, nearest-neighbor pose estimation, and subspace learning to classify Boat images and MNIST digits while comparing PCA and LDA class separation.
- Implemented and evaluated Decision Tree and Random Forest classifiers to predict Titanic passenger survival. Applied preprocessing, feature engineering, information gain, bootstrapping, and ensemble learning to improve classification accuracy through aggregated predictions.