Question on performing statistical fits on large data volume

Hello Rooters,

I have a question regarding how much data one can perform statistical fits on using one of the ROOT packages (e.g. TLinearfitter, TMVA etc.)

Do all the data points on which I want to perform a fit have to fit in the memory ? If yes, then does anyone have a suggestion on what to do if I want to perform fits on data sets much larger than what the memory can hold ?

Thanks a lot.


Hi Arun,

For the TLinearFitter you don’t have to have all the data in memory. Call the function TLinearFitter::StoreData(kFALSE) after the constructor to do this (by default the input data is stored).