Machine Learning: Theory and Implementation

Natanael Alpay · Undergraduate research with Mihaela Vajiac

I studied and implemented classification methods including perceptrons, radial basis function networks, and support vector machines using Python and MATLAB. The project connects their mathematical foundations with experiments and visualizations of learned decision boundaries.

Research poster (PDF)

Visualizing a Decision Boundary

A radial basis function classification experiment showing data points and a learned decision boundary — Decision-boundary visualization from the radial basis function implementation.

These experiments illustrate how the choice of representation affects classification. A single linear threshold perceptron requires linearly separable data for exact classification; problems such as XOR require a richer representation or model.