Statistical Machine Learning for Engineers and Data Scientists. 3 credits, 3 contact hours

Catalog description

Prerequisites: Good knowledge of statistics and probability, as covered in ECE 673 or similar courses, and linear algebra; or permission of instructor. With the explosion of “Big Data” problems, statistical learning has become a very hot field in many scientific areas as well as in marketing, finance, and business. Statistical learning refers to a vast set of tools for understanding data. Specifically, supervised statistical learning involves building a statistical model to predict or estimate an output or pattern based on one or more inputs. For unsupervised learning there are inputs, but no supervising output, and here the goal is to learn relationships and structure from the data. This course provides a systematic introduction to statistical machine learning and pattern recognition using fundamental performance criteria as guiding principles and studies how these principles affect real-world performance. Topics covered include practical techniques as linear and kernel models for classification and regression, unsupervised probabilistic modeling, probabilistic graphical models, and approximate inference, but also theoretical underpinnings as VC dimension and sample complexity.