Open Collections

Some new methods for robust high-dimensional classification

Banff International Research Station for Mathematical Innovation and Discovery

Learning predictive models from genomic data remains challenging due to the high dimensionality and the complexity of the data. I will discuss a few techniques that we have investigated recently to try to overcome some of the challenges: (1) the Kendall and Mallows kernels, which learn a predictive model based on pairwise comparisons between features, and (2) new atomic matrix norms, to learn models with particular sparsity structures such as disjoint support or sparse latent factors.

Mathematics; Statistics; Biology and other natural sciences

video/mp4

Author affiliation: Mines ParisTech

2016-04-20

Attribution-NonCommercial-NoDerivatives 4.0 International

http://hdl.handle.net/2429/57706

Unreviewed

http://creativecommons.org/licenses/by-nc-nd/4.0/