In one embodiment, a device in a network monitors performance data for a first predictive model. The first predictive model is used to make proactive decisions in the network. The device maintains a supervisory model based on the monitored performance data for the first predictive model. The device identifies a time period during which the supervisory model predicts that the first predictive model will perform poorly. The device causes a switchover from the first predictive model to a second predictive model at a point in time associated with the time period, in response to identifying the time period.
Please note there is up to 60 days of latency in this Status indicator for certain status conditions. You can obtain up-to-date Status indicator readings by ordering PAIR for the file.
An application with the status "Published" (which means it is pending) may be recently abandoned, but not yet updated to reflect its abandoned status. However, an application filed less than one year ago is unlikely to be abandoned.
A patent with the status "Granted" may be recently expired, but not yet updated to reflect its expired status. However, it is highly unlikely a patent less than 3.5 years old would be expired.
An application with the status "Abandoned" is almost always current, but there is a small chance it was recently revived and the status not yet updated.
This priority date is an estimated earliest
priority date and is purely an estimation. This date should not be
taken as legal conclusion. No representations are made as to the
accuracy of the date listed. Please consult a legal professional
before relying on this date.