I have run TemporalAnomaly model based on my data. I got actual value from record[ ] , predictive value from result.inferences[“multiStepBestPredictions”][1] and Anomaly score from result.inferences[“anomalyScore”].
I would like to confirm that the “anomalyScore” are calculated by using Predicted and Active columns where the Predicted columns is not equal to predictive value and Active columns also not equal to actual value.
Another question is how can I access the Predicted and Active columns?
Thank you