Alice Smart not present as she has tonsilitis - we will catch her up when she is better.
Michal:
Using appollo/feynman login to HPC cluster.
Running is v. slow, about 12-15 hours for 2M events. Submit to batch farm 2Mx10 overnight
csv format outputs - need to work out how to append these in python
ready to make graphs
Cuts are pT>25 GeV, no eta cut : efficiency ~95% for Hbb and <1% for ggbb
Want to use Gradient-boost BDT in scikit-learn.
Want to learn more about Trigger systems - Lily to google.
Andy:
Goodbye to small angle approx.
Cuts are pT>40 GeV and |eta|<2.5
Gorgeous M/pT versus dR plots for thesis - add chi squared values
Plot of approximate mass for dR<1 - make again for dR<0.6?
Nice update on classifiers
Adjust invariant mass plot to start at 0 on x-axis
Think about the fit: Poisson or Gaussian?
Thomas:
Looking at different scales with fastjet
Random number generator is exponential - pz is problem - is there a better way to generate particle 4 momentum?
Looking at unsupervised learning
Alice Jordan:
Focusing on report as powheg is being a nightmare
Lily and Alice to get powheg under control
Alice Smart not present as she has tonsilitis - we will catch her up when she is better.
Michal: Using appollo/feynman login to HPC cluster. Running is v. slow, about 12-15 hours for 2M events. Submit to batch farm 2Mx10 overnight csv format outputs - need to work out how to append these in python ready to make graphs Cuts are pT>25 GeV, no eta cut : efficiency ~95% for Hbb and <1% for ggbb Want to use Gradient-boost BDT in scikit-learn. Want to learn more about Trigger systems - Lily to google.
Andy: Goodbye to small angle approx. Cuts are pT>40 GeV and |eta|<2.5 Gorgeous M/pT versus dR plots for thesis - add chi squared values Plot of approximate mass for dR<1 - make again for dR<0.6? Nice update on classifiers Adjust invariant mass plot to start at 0 on x-axis Think about the fit: Poisson or Gaussian?
Thomas: Looking at different scales with fastjet Random number generator is exponential - pz is problem - is there a better way to generate particle 4 momentum? Looking at unsupervised learning
Alice Jordan: Focusing on report as powheg is being a nightmare Lily and Alice to get powheg under control