Pattern Recognition and Analysis
MIT OpenCourseWare offers this graduate-level course on characterizing and recognizing patterns in numerical data. Topics include decision theory, statistical classification, maximum likelihood and Bayesian estimation, nonparametric methods, and unsupervised learning and clustering. Applications span user modeling, affect recognition, speech recognition and understanding, computer vision, and physiological signal analysis. The course also touches on active research topics in machine and human learning. Materials include lecture notes and assignments drawn from MIT's Media Arts and Sciences curriculum, taught with a strong theoretical grounding in signal understanding problems. As with other MIT OCW offerings, all course materials are free to access under an open license, with no certificate offered. Suited to students with a background in probability and linear algebra who want a rigorous treatment of pattern recognition fundamentals.