Statistique et Big Data Analytics; Volumétrie, L'Attaque des Clones

This article assumes acquired the skills and expertise of a statistician in unsupervised (NMF, k-means, SVD) and supervised learning (regression, CART, random forest). What skills and knowledge do the statistician must acquire it to reach the "Volume" scale of big data ? After a quick overview of the different strategies available and especially those imposed by Hadoop, algorithms of some available learning methods are outlined to understand how they are adapted to high stresses of Map-Reduce functionalities. The next step will probably be to rewrite them using the R like matricial language which is developped by the communitiesMahout, Spark and Scala. ∗Université de Toulouse – INSA, Institut de Mathématiques, UMR CNRS 5219 †INRA – UR875 MIA-T 1 ar X iv :1 40 5. 66 76 v2 [ st at .O T ] 5 O ct 2 01 4 FIGURE 1 – Armée de clones, baies de serveurs, alignés par milliers dans le hangar d’un centre de données de Google.