This course will focus on several standard computational methods in data sciences and on model selection in statistics. This includes: Concept of statistical model, Supervised model selection, Cross-Validation techniques (Hold-out, V-Fold, Leave-one-out, Leave q-out), Unsupervised model selection, Penalized criteria (AIC, BIC, ICL, slope heuristic), applications of model selection to regression (ridge, lasso), density estimation, model-based classification and clustering, variable selection, dimension reduction (probabilistic PCA).
Supervised and unsupervised model selection, Cross-validation techniques, Penalized criteria, Applications to regression.