I am performing ML tasks. I first trained and tested multiple configurations of BDTs using the TMVAClassification.C macro but when I’m trying to run my model using the TMVAClassificationApplication.C macro, I’m getting only 1 value for the entire dataset. I did make the necessary changes in the application macro.
Hi,
For running your code I would need the .xml files for the weigths. The .C files are to be used to evaluate the method in a standalone C++ code. See for example the TMVA Users Guide paragraph 3.4.
You can try making a tar file with the xml file and upload the tar file, otherwise you can put on some share drive, like dropbox or cernbox or you can try directly by email
Hi Mr. Moneta! Just wanted to follow up with my query. It would be really helpful if you could help me out ASAP since I need to submit my abstract this week.
There is a problem reading the input variables form the input file, since they are not stored as floating type variables. You need to convert to float before evaluating the method and set the correct type in TTree::SetBranchAddress. See my modified macro (you might want to remove some of the debug printout).
I see still a potential issue that the file will contains many inputs equal for many events and this will result in same output values.