SAIP2023

Kernel density estimation and weakly supervised machine learning-based models for Higgs-like signals data classification

6 Jul 2023, 15:40

2h

University of Zululand

Poster Presentation Track B - Nuclear, Particle and Radiation Physics Poster Session 2

Speaker

Nidhi Tripathi (School of Physics, University of the Witwatersrand)

Description

Following the anomaly observed in multi-lepton final states through the decays of heavy scalar resonance in 𝑍𝛾 data at the Large Hadron Collider via proton-proton collisions, we develop a Kernel density estimation-based machine learning model to generate synthetic dataset. The dataset comprises SM Higgs-like signals such as ggF, VBF, WH and ZH. Further we use weak supervised machine learning methods and deep neural network model(s) to classify and discriminate between original and synthetic dataset. We demonstrate the ability of this approach to reproduce the various kinematic observables in the said final states, and preliminary results shows that this model generates the synthetic data reasonably well, where the performance is compared with the standard samples using Monte-Carlo event generators.

Level for award;(Hons, MSc, PhD, N/A)?

PhD