This document outlines four perspectives for identifying patterns in data streams:
1. An identifier/ML model perspective where the general presence of patterns is known or unknown. Known patterns can be found through rule-based or supervised methods, while unknown patterns require unsupervised or rule mining approaches.
2. An analyst perspective where the presence of certain patterns like bots is known or unknown. If their presence is known, rule-based methods can find them. If unknown, hyperparameter tuning may be needed.
3. An observer perspective where an analyst either knows or doesn't know if certain patterns are present. If presence is known but patterns found, unsupervised ML is used. If both presence and patterns are unknown, nothing