1 / 3

Boosted Decision Trees for Supersymmetry Searches

Timothy Brooks, a Year 1 PhD student at Royal Holloway, University of London, explores the application of Boosted Decision Trees (BDT) in supersymmetry searches. The research focuses on distinguishing SU(4) mSUGRA points from background events like t-tbar and W+Z production using BDTs trained on full Monte Carlo samples. The analysis, which uses 400 trees, indicates a significant excess at 31.15 pb^-1 and investigates the effects of over-boosting and background uncertainties. Further studies aim to refine signal extraction in di-lepton and jet multiplicity channels.

hashim
Télécharger la présentation

Boosted Decision Trees for Supersymmetry Searches

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript

  1. Boosted Decision Trees for Supersymmetry Searches Timothy Brooks Year 1 PhD Student Royal Holloway University of London

  2. Boosted Decision TreeSeparation SU4 mSUGRA Point Backgrounds considered: Ttbar production W+Z production Comparisons made to Multi-lepton cuts. Precut requires NLeptons >= 3 Trained on full MC sample (without cuts) to maximise training sample. 400 trees used here. s = 63.4 b = 32.8 Z = 8.96

  3. Boosted Decision TreeSignificance 5 σ excess at 31.15 pb-1 Effect of over boosting investigated. Significance estimation needs to include background uncertainties. Need to move to Di-lepton + 4 jet analysis. Aim is to investigate if a BDT run on control and expected signal regions can extract the top component in the signal region. To recover higher statistics, channels with jet multiplicities lower than 4 will be investigated.

More Related