US 11,657,305 B2
Multi-method system for optimal predictive model selection
Venkata Jagannath Yellapragada, Jersey City, NJ (US); Thomas Hill, Tulsa, OK (US); Daniel Rope, Reston, VA (US); Michael O'Connell, Durham, NC (US); Gaia Valeria Paolini, Canterbury (GB); and Tun-Chieh Hsu, Hoboken, NJ (US)
Assigned to CLOUD SOFTWARE GROUP, INC., Fort Lauderdale, FL (US)
Filed by Cloud Software Group, Inc., Fort Lauderdale, FL (US)
Filed on Jun. 8, 2020, as Appl. No. 16/895,337.
Claims priority of provisional application 62/858,165, filed on Jun. 6, 2019.
Prior Publication US 2020/0387814 A1, Dec. 10, 2020
Int. Cl. G06N 5/04 (2023.01); G06N 20/00 (2019.01); G06N 20/20 (2019.01)
CPC G06N 5/04 (2013.01) [G06N 20/00 (2019.01); G06N 20/20 (2019.01)] 17 Claims
OG exemplary drawing
 
1. An apparatus for generating algorithmic models used for generating predictive analytics from a model data set for a process, the apparatus comprising:
a desirability function module configured to generate a desirability function, wherein the desirability function define:
at least one outcome variable and outcome variable type and at least one predictor variable and at least one predictor variable type; and
at least one algorithmic model accuracy criterion, at least one model analytics type, at least one evaluation criterion for algorithmic model quality, and at least one evaluation criterion for model deployment cost,
wherein the evaluation criterion for model deployment cost includes at least one from a group comprising: cost of scoring the at least one algorithmic model; cost of false-positive prediction per categorical outcome; cost of false-negative prediction per categorical outcome; value of correct prediction per categorical outcome; cost for prediction error per continuous outcome; cost of acquiring data for each predictor variable; and cost of model building and recalibration, and wherein the cost of false-positive prediction per categorical outcome is stratified by each input value per class; and
an automated machine learning module configured to:
generate at least one algorithmic model having a variable set selected according to the desirability function; and
train the at least one algorithmic model against the model data set.