Thijs
After comments yesterday, now using likelihoods instead of a hit coincidence count with specific cuts. I'm using two methods to find the two bangs:
1.) Fix neutrino vertex and vary location of second shower vertex.
2.) Fix second shower vertex and vary location of neutrino vertex
Dorothea: which energies do you assign to the two showers?
Thijs: They are both assigned equal energy.
Dorothea: Robert's studies seemed to indicate that the success of a double bang algorithm depends largely on the number of hits belonging to each individual shower.
Bouke: Are the vertices fixed to the MC truth?
Thijs: Yes.
The tau propagation length is varied in steps of 1-2 m.
In accordance with this, the algorithm finds the maximum likelihood for the location of the second vertex at typically 1-5m distance from the MC truth.
For some events the maximum likelihood is found very far away from the MC-truth.
This seems to happen a bit more often for the case where the decay vertex is fixed.
Dorothea: What event weighting is used? E.g. what is the spectral index?
Ronald: Usually 1.4.
Thijs: If this is true, then this is also the case in my generated events, since I didn't change anything.
Dorothea: So this means that you are incorporating more high-energy events than you would expect in-situ.
Ronald: What is the longer term plan for this?
Thijs: The steps I have shown now will be part of some prefit. The end-goal is a full likelihood scan over all event parameters (energy, vertex, direction).
Ronald: What minimizers are you using?
Thijs: Currently using standard scipy and numpy minimizers. I would like to look into other minimizers in the future.
Brían
Action points for track reconstruction documentation have almost all been addressed. Two points still need looking into.
Bouke
A git project has been set up where we can collect any points of action discussed during the reconstruction meetings. Link will be sent around after the meeting. For now, I have updated the project with a git issue revolving about the points that were discussed during our meeting which focussed on shower reconstruction.