Coursera Flash Sale
40% Off Coursera Plus for 3 Months!
Grab it
Learn anomaly detection fundamentals through this comprehensive lecture covering three key aspects: understanding expectations for anomaly detection, identifying the shape of interesting subsets to consider, and evaluating whether a discovered subset is genuinely interesting. Explore the mathematical foundations including likelihood definitions from normal independent and identically distributed examples, and master the log-likelihood ratio calculation as ln(L(S) * L(X\S) / L(X)). Discover techniques for finding interval subsets that maximize the log-likelihood ratio in change point models, and develop skills in quantifying results using permutation tests and p-values for statistical validation of detected anomalies.
