Which of the following items apply to anomaly detection? (Choose all that apply.)

A. Use AD on KPIs that have an unestablished baseline of data points. This allows the ML pattern to perform it’s magic.

B. A minimum of 24 hours of data is needed for anomaly detection, and a minimum of 4 entities for cohesive analysis.

C. Anomaly detection automatically generates notable events when KPI data diverges from the pattern.

D. There are 3 types of anomaly detection supported in ITSI: adhoc, trending, and cohesive.

Explanation:
The KPI must be split by entity, and a minimum of four entities is required.

If the KPI diverges from the normal pattern, ITSI creates a notable event in Episode Review.

Reference: https://docs.splunk.com/Documentation/ITSI/4.10.2/SI/AD