User Tools

Site Tools


~June 6th - 17th: Tried to find an analytical best-fit function for inv_JetJet histogram to no avail. Decided to try numerical approach.

~June 20th: First attempt at peak search script by comparing slopes. Success, but too sensitive.

~June 27th: With help of Dr. Chekanov, implemented script with finds peaks by comparing the average value of the slope up to the point in question. Works very well

June 27th onwards: 1)implemented linear regression script with creates a background for each peak found by linear regression of the first and last points of the peak. 2)added script which compares the original peak with the linear regression and calculates the statistical significance of the peak

July 12th: Was able to implement code that locates possible peaks, forms a background linear regression of the peak, and then calculates the statistical significance of the peak by comparing to the regression. However, some statistical significance values are higher than they should be (easily seen by eye) so there must still be a bug of some sort.

Report: non-parametric_peak_finder_for_general_searches_in_atlas_data.pdf

people/merickson.txt · Last modified: 2013/05/30 18:53 (external edit)