[PFR] Expanded ML-driven anomaly detection for CloudTrail

[jonrau1]

Story
As a user of SyntheticSun I want to utilize additional machine learning algorithms to detect anomalies in my CloudTrail logs based on different features so that I can reduce the residual risk of any unauthorized access

Definition of Done

• Updated architecture, maybe another Appendix
• Updated readme and permissions
• Expanded CT Training Grounds in Appendix A to include new Dockerfile, helper scripts and subdirectory
• ML that works?

Nice to Have

• Not being an idiot

Additional Information

• https://docs.aws.amazon.com/sagemaker/latest/dg/randomcutforest.html
• https://docs.aws.amazon.com/sagemaker/latest/dg/k-nearest-neighbors.html