Model choice and diagram review

Model choice and diagram review

animation
iiot
maintenance
TipAda, Applications and calculation-audit guideCheckpoint: Model Choice

You now know:

  • Supervised learning can answer whether a bearing will fail in the next 30 days, but it needs labeled failure examples.
  • Unsupervised anomaly detection can start from normal operation only, but it still needs engineer review before action.
  • RUL and time-series models estimate time to failure only for the asset class, fault mode, and operating context represented in the training data.
NoteLabel the Diagram

Back to the chapter