En naviguant sur notre site vous acceptez l'installation et l'utilisation des cookies sur votre ordinateur. En savoir +

Menu Logo Principal AgroParisTech

MIA Paris

Atelier Happy R

25 janvier 2019, 14h-17h - Salle Actia, AgroParisTech

Efficient analysis of large-scale matrices with two R packages: bigstatsr and bigsnpr, par Florian Privé.

Florian Privé

R package bigstatsr provides a special class of matrix whose data is stored on the disk instead of the RAM, but you can still access the data almost as if it were in memory. It is particularly useful is you have a large matrix to analyze but not enough RAM on your computer. It can still be useful for matrices that fit in your RAM because package bigstatsr provides very efficient and parallelized algorithms (have you ever found cor or svd too slow?). I will present the statistical and helper functions that are provided by package {bigstatsr} for this kind of matrices. R package bigsnpr, on top of bigstatsr, provides some tools that are specific to the analysis of genetic data. We’ll see what I can predict from your DNA using these two packages.

https://stateofther.github.io/talk/2018-01-25-bigstatsr/