MSDS 411-DL Unsupervised Learning Methods.

This course introduces traditional and modern methods of unsupervised learning. Students see how to represent relationships among many continuous variables using principal components and factor analysis. They identify groups of individuals and groups of variables with cluster analysis and block clustering. They explore relationships among categorical variables with log-linear models and association rules. They visualize multivariate data with lattice displays, multidimensional scaling, and t-distributed stochastic neighbor embedding. And they detect anomalies using autoencoders and probabilistic deep learning. This is a project-based course with extensive programming assignments. Prerequisites: MSDS 400-DL Math for Modelers and MSDS 401-DL Applied Statistics with R.

Back to main page for Analytics and Modeling.