Plans

[arogozhnikov]

• have a sample (finally!!!)
• restructure code to work with new data format
• so we need event building? (I think, yes)
• address memory issues
• RANSAC? Or anything else to find pz, pt and entry point

  much later

• think about moving BDT to C++

[ewengillies]

• have a sample (finally!!!)
• restructure code to work with new data format DONE
• so we need event building? (I think, yes) DONE
• address memory issues MOSTLY DONE

done

• 10,000 background bunches built
• 40,000,000 signal tracks simulated
• 5,000 more background bunches still to build, but IO is causing problems.

now

• First draft of paper
• Fix IO issue

very soon (for paper)

• Figure out how to estimate the error/performance of our hit filter
  • Review previous literature.
  • Application of Kalman filtering to track and vertex fitting, 1987
  • Low-momentum track finding in Belle II, 2012
  • Demonstrator of the Belle II Online Tracking and Pixel
  • Forward Tracking in the ILD Detector
  • Current ideas :
  • Check momentum at each hit point before and after hit filtering (weighted histogram in terms of prediction score, or cut on prediction score)
  • Check position distribtutions (as with momentum).
• Optimise parameter space against our current data set
  • I want to scan. I have 500 cores at imperial to work with. This is doable.
• Try with x2 bunch size (double the background) to show robustness against background.

later

• RANSAC? Or anything else to find pz, pt and entry point

much later

• Move algorithm to C++
  • BDT = easy (probably)
  • Reweighted inverse hough = harder