Read different size of vector using TMVA macro

[Hokyeong]

Dear experts,

I tried modify the TMVAClassification.C macro to analysis for single electron data
and I get the following message.

: Dataset[dataset] : ERROR while preparing training and testing trees:
: Dataset[dataset] : multiple array-type expressions of different length were encountered
: Dataset[dataset] : location of error: event 0 in tree Signal_Tree of file SingleElectron.root
: Dataset[dataset] : expression track_R has Dataset[dataset] : 3 entries, while
: Dataset[dataset] : expression track_Eta has Dataset[dataset] : 6 entries
: Dataset[dataset] : Need to abort
***> abort program execution
terminate called after throwing an instance of ‘std::runtime_error’
what(): FATAL error

The data file I used have variable which is consist of vector of different size.
According to error message, that fact seems to be causing error.

Vector entries in TMVAClassification and Application

Hello dear ROOTforum, I have trouble with TMVA’s macros. The tree source has branches type vector, in the part of dataloader I put this dataloader->AddVariable( "B_J_px2", "B_J_px2", "units", 'F' ); and works fine. But in the part of Application, I stated the same variables in the part of classification (like tutorial explain) TMVA::Reader *reader = new TMVA::Reader( "!Color:!Silent" ); // Create a set of variables and declare them to the reader // - the variable names MUST correspon…

I saw the above topic on ROOT forum.
And found the something I want. In the end of topic kialbert saied “If the vector is of variable size we need to consider another approach.”
But I have no idea how the problem can solve.
Please give me some advice.

I attached the macors.
TMVAClassification.C (3.1 KB)

And this is the data file I used

SingleElectron.root

Background.root

Thank you a lot for your help in advance.

Best regards,
Hokyeong

[kialbert]

Hi,

Could you elaborate on why the two variables have different length? In a single event there should be a set number of tracks, and all variables referring to track data should have an entry for each track.

I don’t see why the branch track_R should have only 3 entries while track_Eta should have 6 entries.

The most straight forward solution is to reformat your data so that each branch corresponds to a TMVA variable.

Cheers,
Kim

[dlin]

Dear Kim,

I am learning to use the TMVA classification and facing the similar problem. I have a vector branch with different size for different events, I am wondering if it’s possible to add this branch into the dataloader? and how can I do this?

Thank you very much in advance!
Best regards,
Dexu

[kialbert]

Hi,

(In the future, please post a new question instead of reviving an old one. This to help keep better order in the forum. )

Unfortunately, the answer is a negative: This is not something TMVA is designed to handle. You can add array elements (e.g. dataloader->AddVariable("myArrayBranch[0]")), but not variable length arrays.

Cheers,
Kim

[dlin]

Many thanks for the reply!
I will keep in mind to post as a new thread.