Applied & Computational Mathematics seminar: Optimal Time-Dependent Classification for Diagnostic Testing

Speaker:  Prajakta Bedekar, NIST

Title:  Optimal Time-Dependent Classification for Diagnostic Testing

Abstract:  Antibody tests can identify past infection by quantifying the immune response of an infected individual, thereby providing guidance for decisions about public health measures. The relative antibody measurements change with time due to the variation in an individual's antibody levels and prevalence in the population as the pandemic progresses. In this talk, I will demonstrate the use of optimal decision theory to develop a time-dependent, probabilistic classification scheme which takes both the personal and the population-level effects into account. These classification domains change with time and suggest a natural adaptive scheme for estimation of prevalence, considering the progression of the pandemic. I will illustrate the results by using a combination of SARS-CoV-2 and synthetic data sets, and detail the type of data needed to execute this scheme in real-world settings. This talk is based on joint work with Paul Patrone and Anthony Kearsley.

Time:  Friday, November 18, 2022, 1:30pm-2:30pm

Place:  Exploratory Hall, Room 4106 or Zoom