Improving COVID-19 Testing Efficiency using Guided Agglomerative Sampling

Fayyaz Minhas; Dimitris Grammatopoulos; Lawrence Young; Imran Amin; David Snead; Neil Anderson; Asa Ben-Hur; Nasir Rajpoot

In order to evaluate the efficacy of this approach, we constructed a simple simulation in which individuals are assigned random test labels ( = 1 with probability and = 0 with probability 1 − ). Each individual is then assigned a degree of belief . We tested with both a random degree of belief (no belief information) and varying degrees of belief as measured by the concordance between and by using an additive normal distribution noise prior = + (with ~ℵ(0, )) with the degree of noise controlled by the standard deviation parameter . For a given simulation setting (number of individuals, prior probability and belief control factor ), we calculate the number of required tests by the proposed sampling method. In order to get reliable statistical estimates of the distribution of the number of required tests for a given simulation setting, we repeated the simulation multiple times with the same setting and plotted the distribution of the number of required tests using a box plot.