Anomaly Detection — Scores and identifies observations that deviate meaningfully from a model of normal behaviour or data distribution.
Scores and identifies observations that deviate meaningfully from a model of normal behaviour or data distribution.
Anomaly is not synonymous with rarity; the model of normal behaviour and the decision threshold must account for false-alarm cost and changing data.
Publications whose title or summary directly references this concept.
Visual anomaly detection through models of normality, anomaly scores, temporal context, false-alarm control and distribution shift.
Acoustic anomaly detection through models of normal operation, transient and long-term deviations, trend analysis and multi-sensor validation.
Course notes connecting SVMs, k-NN, K-Means, Naive Bayes, decision trees, regression, PCA/LDA, HMMs, n-grams, anomaly detection, evaluation, and model selection within one statistical-learning framework.
Uses only this site's published content. How it works
