| CPC G06N 5/04 (2013.01) [G06N 20/00 (2019.01); G07C 5/008 (2013.01); G07C 5/085 (2013.01); G06F 3/0482 (2013.01)] | 20 Claims |
|
1. A method of predicting maintenance events, the method comprising:
determining a training data set comprising:
a first listing of components removed from service due to an unplanned maintenance event;
performance data for each component on the first listing; and
operational data regarding maintenance for each component on the first listing;
determining, using one or more machine learning models, one or more test listings of components predicted to be removed from service based on the training data set;
wherein each of the one or more machine learning models is associated with one of the one or more test listings;
identifying, based on a comparison of the one or more test listing to the first listing, false positives or false negatives in each of the one or more test listings;
determining, based on the identified false positives or false negatives, a weight for each of the one or more machine learning models, the weight associated with a relative value of each of the one or more machine learning models within a prediction module; and
generating, by the prediction module, a second listing of components from a plurality of components in service,
wherein each component on the second listing of components is predicted to have an unplanned maintenance event within a predetermined period of time, and
wherein the weight is determined to reduce a number of false positives or false negatives in the second listing of components.
|