Stats 102B Computation and Optimization for Statistics
Course
description:
Introduction to computational methods and optimization useful for statisticians.
Use of computer programming to solve statistical problems. Main topics include:
1) Matrix Algebra: vector and matrix computation, connections to statistics, eigenvalue decomposition.
2) Principal Component Analysis: multivariate normal distribution, principal components,
dimension reduction, principal component regression.
3) Differentiation and Optimization: gradient and Hessian, Newton's method, KKT theory and constrained optimization,
penalized least squares, coordinate descent.
4) EM and MM algorithms: missing data, the EM algorithm, the MM algorithm, examples.
Link to the course CCLE site for lecture notes, homework assignments, etc.