US 12,406,024 B2
Balance weighted voting
Lei Tian, Xian (CN); Han Zhang, Xian (CN); Ning Zhang, Xian (CN); Xiao Li Zhang, Xian (CN); Yi Shao, Xian (CN); Jing Xu, Xian (CN); and Xue Ying Zhang, Xian (CN)
Assigned to International Business Machines Corporation, Armonk, NY (US)
Filed by INTERNATIONAL BUSINESS MACHINES CORPORATION, Armonk, NY (US)
Filed on Dec. 13, 2021, as Appl. No. 17/643,877.
Prior Publication US 2023/0185882 A1, Jun. 15, 2023
Int. Cl. G06F 18/21 (2023.01); G06N 20/20 (2019.01)
CPC G06F 18/2193 (2023.01) [G06N 20/20 (2019.01)] 15 Claims
OG exemplary drawing
 
1. A method comprising:
receiving a scoring request;
generating, by a machine learning model and based on the scoring request, a list of performance factors and corresponding performance thresholds for selecting a subset of component models of a plurality of component models to include within an ensemble model for responding to the scoring request;
selecting the subset of component models from the plurality of component models by determining that historical performance of respective ones of the component models of the subset of component models satisfy the performance thresholds corresponding to the performance factors within the list;
generating a plurality of scores using the subset of component models;
normalizing the plurality of scores;
calculating an evaluation-based weighting factor from a first subset of the normalized scores;
calculating a prediction-based weighting factor from a second subset of the normalized scores;
calculating a balanced weighting predictor from the evaluation-based weighting factor and the prediction-based weighting factor; and
returning the balanced weighting predictor as an ensemble score for the scoring request.