Exam AWS Certified Machine Learning - Specialty topic 1 question 189 discussion

AD are the right answer

Random Cut Forest (RCF) (Option D): RCF is an unsupervised algorithm designed for anomaly detection. It can identify unusual patterns in the data without requiring labeled examples of fraudulent transactions1. IP Insights (Option A): IP Insights is another unsupervised algorithm that can detect anomalies based on IP address usage patterns. It is particularly useful for identifying suspicious activities related to IP addresses

A. YES - IP Insights works unsupervised on IP addresses; builtin algorithm B. NO - k-NN is unsupervised clustering, does not help with anomalities C. NO - Linear learner is supervised D. YES - Random Cut Forest (RCF) is unsupervised anomalities E. NO - XGBoost is supervised

IP Insights is an unsupervised learning algorithm that learns the usage patterns of IP addresses. It can capture associations between IP addresses and various entities, such as user IDs or account numbers. It can also identify anomalous events, such as a user attempting to log in from an unusual IP address, or an account that is creating resources from a suspicious IP address1. Random Cut Forest (RCF) is another unsupervised algorithm for detecting anomalous data points within a dataset. It can handle arbitrary-dimensional input and scale well with respect to number of features, data set size, and number of instances. It can detect anomalies such as unexpected spikes in time series data, breaks in periodicity, or unclassifiable data points2.

Can't be A, as we don't have data in the format expected for IP Insights algorithm(https://docs.aws.amazon.com/sagemaker/latest/dg/ip-insights-training-data-formats.html).

A and D are correct. A. IP Insights for Pattern recognition. D. Random Cut Forest (RCF) for Anomaly detection B,C,E are normally Supervised learning algorithm which are against the wordings "There is no label ..."

C is not part of the answer. IP insight because the data contain IP address. RCF because the data is unlabeled and anomaly is being detected for fraud.

AD are correct

apprently AD

AC is the correct answer to detect anomalies