Issue with MVA and NaN Values

TMVA
1

Hello everyone,

I’m facing a problem with my code for limit calculations. I have a dataset with variables m, pt, eta, and phi, I’m applying some BDT scores from TMVA training.
The issue I’m encountering is as follows:
: 2-th variable of the event is NaN → return MVA value -999,
: that’s all I can do, please fix or remove this event.
I’m getting this error for different variables in different events. It seems that some of my variables are becoming NaN, and this is causing the MVA value to be set to -999. This behavior is unexpected and is affecting the accuracy of my calculations.
i have scanned all my root file and non of them have a NAN value.

If anyone has encountered a similar issue or has any insights into why this might be happening, I would be incredibly grateful for your help.

Thank you in advance for your assistance!

2

Hi @asmaa.aboulhorma , great to see you on the forum.

I am not sure how large your data set is but a fast way to check
input data is to calculate the mean and variance of each column. If those
numbers are finite, none of your input will be NaN.

-Eddy

3

hi @Eddy_Offermann, Thank you for your feed back.
I have checked my tree by scanning each column, and it seems to be normal. I also thought that maybe the problem was because my tree had double and the training used float. I just converted my tree variable to float but still have the same problem.

4

Did you scan by looking at the values or doing an operation with them ?

5

I scaned by looking looking at the values

6

hi @Eddy_Offermann

I have calculated the mean and variance of each column. all of those numbers are finite. i have attached a screen of one of the outputs.

image
image924×652 250 KB