diff --git a/ml/README.md b/ml/README.md index 2578993e2e..f6fd923ab5 100644 --- a/ml/README.md +++ b/ml/README.md @@ -248,8 +248,6 @@ In terms of anomaly detection, the most interesting charts would be the `anomaly - `anomaly_detection.dimensions`: Percentage of anomalous dimensions. - `anomaly_detection.detector_window`: The length of the active window used by the detector. - `anomaly_detection.detector_events`: Flags (0 or 1) to show when an anomaly event has been triggered by the detector. -- `anomaly_detection.prediction_stats`: Diagnostic metrics relating to prediction time of anomaly detection. -- `anomaly_detection.training_stats`: Diagnostic metrics relating to training time of anomaly detection. Below is an example of how these charts may look in the presence of an anomaly event.