LFU:online

Regression modelling with large data sets

- Linear and generalized linear regression models: model definition, estimation and inference

- Basic regression algorithms: least squares problem and solutions, complexity and memory, iterative re-weighted least squares

- Tall data (massive $n$): incremental bounded-memory algorithms, stochastic gradient descent, parameter elimination, inference, software and tools, analysis of real datasets

- Sparsity: sparsity-inducing penalties, lasso regression, proximal Newton algorithms, cross-validation, applications in generalized linear models, software and tools, analysis of real datasets

The course aims at offering students an understanding of the core theory and methods that can be used for learning regression models where there is an abundance of information either in terms of available observations or in terms of available explanatory variables.

Take home assessment

- Golub, G. H. and C. F. Van Loan (2012). Matrix Computations (4th Ed.). Johns Hopkins University Press.

- Hastie, T., R. Tibshirani and J. Friedman. The Elements of Statistical Learning (2nd Ed.). Springer.

- T. Hastie, R. Tibshirani and M. J. Wainwright (2015). Statistical Learning with Sparsity: The Lasso and Generalisations. Chapman and Hall/CRC Press.

First year Master courses, familiarity with R

Kurstermine: 30.01.2017: 14.30 - 16.00 im SR 7 (Sowi); 31.01.2017: 14.30 - 16.00 im Besprechungsraum (Statistik); 1.2.2017: 14.30 - 16.00 im SR 7 (Sowi)