Search for the Top Quark at D0 using Multivariate Methods

TL;DR

This paper reports on the search for the top quark at Fermilab using multivariate analysis techniques, including neural networks and H-matrix methods, to identify top events and measure production cross-sections.

Contribution

It introduces multivariate methods such as neural networks and H-matrix analysis to improve top quark detection in proton-antiproton collision data.

Findings

Identification of a candidate top event with high likelihood.

Measurement of top-antitop production cross-section as 6.7 +/- 2.3 pb.

Consistent results across different multivariate analysis methods.

Abstract

We report on the search for the top quark in proton-antiproton collisions at the Fermilab Tevatron in the di-lepton and lepton+jets channels using multivariate methods. An H-matrix analysis of the e-mu data corresponding to an integrated luminosity of about 13.5 pb-1 yields one event with a likelihood to be a top event (assuming top mass of 180 GeV/c**2) that is 10 times more than WW and 18 times more than Z -> tau tau. A neural network analysis of e+jets channel with about 48 pb-1 of data shows an excess of events in the signal region and yields a cross-section for top-antitop production of 6.7 +/- 2.3(stat.) pb, assuming a top mass of 200 GeV/c**2. A PDE analysis of e+jets data gives results consistent with the above.