A MadEvent AnalysisProgram

I've written a program that contains a "reconstructing" math algorithm that histograms kinematic and other interesting variables for MadEvent roottuples.

First, let's talk about how to make MadEvent roottuples. Go to http://madgraph.hep.uiuc.edu/ and register. Input the process which you would like to generate events for. For example, gb>tw->e-ve~w+b gives the Wt channel for single top decay with the primary W decaying into an electron and neutrino. Check that you made the diagrams by selecting the "Process Information" link. Next, Click "Code Download" and save the tar file.

Untar the file and open the run_card.dat file in the Cards folder. Name the run (run_tag) and put the desired number of events (nevents). The default is "fermi" for the run name and 10,000 for the number of events. Save the run_card.dat file. Go the main folder and type bin/generate_events

Choose 0 for a serial run and choose (usually the same) run name. The code you downloaded will then generate the events you wish. It outputs zipped event files.

Go back now to http://madgraph.hep.uiuc.edu/ and choose the Tools tab. Choose Decay Interface. Here you can upload the event file you just generated and force a decay. For example, I decayed the secondary W to jet-jet. Indicate that you want a root output for the decay.

Once you have the decayed root file, you're ready to use my code. As of May 2008, the code is stored at /home/lareseda/madgraph/twminus/jjwplusb/new/decayout/newratio.C. You will need to edit the code to run it on your own files. Currently it inputs one file (f1) and chains together two other files (f2 and f3) and then histograms all the variables for both. For example, it histograms pt_l.gif for the transverse momentum of the lepton in f1 and pt_l2.gif for the transverse momentum of the lepton in the chain. Then bin-by-bin it takes the ratio of f1 to f2 and f3. This is saved as pt_lratio.gif. Finally, the program saves a rootfile of all the histograms produced, histo.root. Also note that by default the program saves all the plots in a plot/ directory that must previously be instated from where the program itself is saved.

The program needs some improvement: First, it does not differentiate between the first jet (highest pt) and the second jet (second-highest pt). It would be interesting to implement this differentiation. Second, the program attempts to make two di-plots. Examining the ratio of di-plots would be great, but I didn't figure out how to take the ratio of them. The program produces the two di-plots successfully but the ratio of the two plots is empty. One quick improvement might be to use a Lego plot instead of a scatter. It's not clear that taking the ratio of two scatter plots is even a good idea.

The purpose of writing this program was to show that the generation of the Wt channel for single top may be inappropriate. During the generation, the primary W was forced to decay to lepton and neutrino, while the secondary W decayed to jet-jet. However, it is equally likely that the opposite will occur: the primary W -> jj and the secondary W -> lepton, neutrino. The file indicated above examines the ratio of what is "now" and "what should be." Preliminary results show that there is noticeable differences between what is used now and what should have been generated.

Here's a breakdown of my directory structure of files relating to MadGraph.

Starting from /home/lareseda/ there is the folder madgraph. Inside madgraph are two folders, tantiwplus and twminus. These represent the two basic charge schemes available for this decay. Inside twminus (where most of my work was done) are folders named eveantiwplusb and jjwplusb. These names are what the primary W decays into. The most work was done inside the jjwplusb folder. Opening the jjwplusb folder, you see all the files that come inside the tar package when you do the bin/generate_events. I generated 10,000 events for these files. Also there is a file marked "again." This is a folder I made as troubleshooting. It contains 20,000 events but seems to be faulty in some way. The folder marked "new" contains a root file with 20,000 events that works well.

That is, /home -> lareseda -> madgraph -> tantiwplus , ( twminus -> eveantiwplusb, jjwplusb )

-- DanielleLarese - 01 May 2008

This topic: Computing > WebHome > AnalysisProgram
Topic revision: 16 Oct 2009, TomRockwell
This site is powered by FoswikiCopyright © by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
Ideas, requests, problems regarding Foswiki? Send feedback