US 11,657,296 B1
System and method of portfolio matching
Bradley Philip Burns, Arlington, VA (US); Anthony Dean Bruce, Washington, DC (US); and Arvind Bhusnurmath, Alexandria, VA (US)
Assigned to APPLIED PREDICTIVE TECHNOLOGIES, INC., Arlington, VA (US)
Filed by APPLIED PREDICTIVE TECHNOLOGIES, INC., Arlington, VA (US)
Filed on Mar. 10, 2021, as Appl. No. 17/197,523.
Application 17/197,523 is a continuation of application No. 13/754,039, filed on Jan. 30, 2013, granted, now 10,949,752.
Int. Cl. G06N 5/02 (2023.01); G06Q 10/06 (2023.01)
CPC G06N 5/02 (2013.01) 20 Claims
OG exemplary drawing
 
1. A computer-implemented method comprising:
calculating, by a computer, a performance trend for a cohort dataset comprising data associated with a subset of control locations of a set of control locations;
iteratively adjusting, by the computer, the cohort dataset by:
in response to detecting a first deviation between the performance trend of the cohort dataset and a performance trend of a test location, dynamically adjusting, by the computer, the cohort dataset, such that data associated with at least one control location within the cohort dataset is dynamically replaced by data associated with at least one additional control location from the set of control locations that was not included in the cohort dataset;
detecting, by the computer, a second deviation between a performance trend of the adjusted cohort dataset and the performance trend of the test location,
wherein the computer continues to iteratively adjust the cohort dataset until the first and second deviations satisfy a threshold; and
executing, by the computer, a test versus control analysis protocol using the adjusted cohort dataset.