HBM 538E - Mathematical Methods in Data Analysis & Machine Learning

Course Objectives

1. To teach mathematical backgrounds of data analysis and machine learning methods.
2. To perform computational analysis of data analysis and machine learning algorithms.
3. To select the appropriate method for the given problem and apply this computational method in a computer environment efficiently.
4. To examine and compare the results elicited from the computational methods

Course Description

Matrix spaces, matrix factorization, eigenvalues and eigenvectors, singular value decomposition, Eckart-Young Theorem, vector and matrix norms, principal component analysis, least squares method, linear equation systems, exponential matrices, derivatives of matrices, saddle points, minmax problem, function minimization, gradient descent method, stochastic gradient descent method, artificial neural networks, back-propagation algorithm, partial derivatives, convolutional neural networks, learning function, finding clusters in graphs