KI-Glossar
Das vollständige Wörterbuch der Künstlichen Intelligenz
Classical Statistical Methods
Approaches based on statistical tests and probability distributions for identifying outliers.
Isolation Forest Detection
Ensemble algorithm using random decision trees to effectively isolate anomalies.
Autoencoders for Anomalies
Neural networks learning to reconstruct normal data, with anomalies having high reconstruction errors.
One-Class SVM
Support Vector Machine learning a boundary around normal data to detect outliers.
Time Series Detection
Specialized technodes for identifying anomalies in sequential and temporal data.
Density-Based Methods
Algorithms such as DBSCAN and LOF identifying anomalies as points in low-density regions.
Streaming Detection
Real-time algorithms processing continuous data streams to detect dynamic anomalies.
Contextual Anomalies
Detection of abnormal observations only within a specific context or given environment.
Multivariate Detection
Techniques analyzing relationships between multiple variables to identify multidimensional anomalies.
Clustering Methods
Approaches identifying anomalies as points not belonging to any cluster or being far from centroids.
Collective Anomalies
Detection of groups of observations that are abnormal together but not individually.
Deep Learning Detection
Use of deep neural networks (GAN, LSTM, Transformers) for complex anomaly detection.