36-736 Methods for Correlated Data and Bayesian Statistics 2

6 units, Carnegie Mellon University

This course will cover methods for spatially and temporally correlated data as well as unsupervised methods for clustering and dimension reduction. Examples include Gaussian processes, Kalman filters, finite mixture models, hidden Markov models, and probabilistic factor models. We will cover topics such as rejection sampling, variational inference and, time permitting, sequential Monte Carlo particle filters. Students will conduct data analyses using probabilistic programming languages as well as implementing a subset of methods from scratch. Prerequisite Methods for correlated data and Bayesian statistics 1. Familiarity with python will be helpful.

Prerequisites: none listed

Courseweave needs JavaScript to show sections, workloads and planning tools.