Abstract: A method of behavior monitoring includes determining, by one or more trained behavior models associated with a monitored asset, output data indicative of operation of the monitored asset. The method also includes determining a risk score based on the output data and determining feature importance data based on the output data. The method further includes determining whether to generate an alert based on the risk score and the feature importance data.

Abstract: A method includes providing behavior model input data to a behavior model to generate behavior model output data and generating an initial state prediction based on the behavior model output data. The method also includes, based on the initial state prediction indicating a particular state, generating state prediction statistics based on the initial state prediction and historical state predictions indicating the particular state. The method further includes providing classifier input data to a classifier model to generate a classification output indicating whether the initial state prediction is reliable. The classifier input data generated based on the state prediction statistics.