Sign-in Links
STAT 462 Statistical Learning I Units: 3.00
A working knowledge of the statistical software R is assumed. Classification; spline and smoothing spline; regularization, ridge regression, and Lasso; model selection; treed-based methods; resampling methods; importance sampling; Markov chain Monte Carlo; Metropolis-Hasting algorithm; Gibbs sampling; optimization. Given jointly with STAT 862.
Learning Hours: 120 (36 Lecture, 84 Private Study)
Offering Faculty: Faculty of Arts and Science
Course Learning Outcomes:
- Apply Markov Chain Monte Carlo for approximating the posterior distributions in Bayesian statistical Analysis.
- Implement common algorithms in R for simulating random variables/vectors from standard and non-standard distributions.
- Use standard Monte Carlo methods and importance sampling for approximating integrals, expectations and probabilities.
- Understand common unsupervised learning methods including density estimation, clustering and dimension re-duction techniques.
- Understand the EM algorithm and its implementation in estimation for mixture models and censored data.
- Understand the use of spline and penalization methods in supervised learning.