MachineLearningProject(from scratch) and HandwrittenDigitClassifier model training
Other
The project implemented multiple machine learning models and training pipelines from scratch using Python and numerical libraries. It focused on transforming raw inputs into learned parameters for classification and regression tasks. No human annotation workflow is described, so labeling-related work is inferred as dataset preparation and model training evaluation. • Implemented Linear Regression and Polynomial Regression using gradient descent and normal equations. • Built Logistic Regression and an N-layer Neural Network with backpropagation. • Implemented Softmax regression and K-Means clustering with visualization support. • Developed a handwritten digit classifier using a neural network trained on 28x28 pixel images.
2025 - Present