US 11,810,013 B2
Systems and methods for alerting to model degradation based on survival analysis
Weichen Wang, Cary, NC (US); Eliza Salkeld, Raleigh, NC (US); Shuyan Lu, Durham, NC (US); and Shanna Hayes, Cary, NC (US)
Assigned to INTERNATIONAL BUSINESS MACHINES CORPORATION, Armonk, NY (US)
Filed by International Business Machines Corporation, Armonk, NY (US)
Filed on Nov. 14, 2019, as Appl. No. 16/683,633.
Prior Publication US 2021/0150396 A1, May 20, 2021
Int. Cl. G06N 7/01 (2023.01); G06N 20/00 (2019.01)
CPC G06N 7/01 (2023.01) [G06N 20/00 (2019.01)] 12 Claims
OG exemplary drawing
 
1. A computer-implemented method for alerting to analytical model degradation in a data processing system comprising a processing device and a memory comprising instructions which are executed by the processing device, the method comprising:
obtaining, by the processing device, metric data associated with an analytical model;
performing, by the processing device, a model metric distribution analysis of the analytical model by evaluating the location of the metric data within a calculated statistical distribution of metrics of other models;
determining, by the processing device, a time period at which the model was trained based on the obtained metric data;
identifying, by the processing device, a survival time period of the model based on both the determined time period at which the model was trained and a failure time period of the model;
repeating the above steps to determine a survival time period for the plurality of analytical models;
performing, by the processing device, a survival analysis based on the survival time period for the plurality of analytical models;
outputting, by the processing device, a training notification indicating that the given model is required to be retrained based on the survival analysis; and
retraining the analytical model.