Séminaire de Statistique et Optimisation

Learning linear models with missing values: a blessing or a curse?

by Claire Boyer (LPSM, Sorbonne Univ.)

Amphi. Schwartz (1R3)

Amphi. Schwartz



Missing data is becoming ubiquitous in machine learning, with this problem growing as data volumes increase. 
We will take a (brief) look at how data science deals with missing data. And we will try to shed some theoretical light on widely-used practices. To do this, we will focus on the seemingly simple problem of linear regression.