MGS 662LEC - Optimization Methods for Machine Learning

Investment in government and business infrastructure has lead to the accumulation of vast amounts of data in recent years. This course will discuss how techniques from convex optimization can be used to extract useful knowledge and business value from the data collected. It introduces students to the theory of convex optimization of relevance to managerial decision making and machine learning. Topics include convex sets and functions, formulation of convex optimization problems, and convex optimization algorithms including gradient, sub-gradient, proximal and interior point methods. Numerous examples will be chosen from machine learning problems including classification, regression and clustering. Students will have hands on experience with the R programming language and optimization packages including MOSEK. We will examine real world examples and case studies from text mining, medical applications, fraud detection, finance, and social networks.

Credits: 3