Słownik AI
Kompletny słownik sztucznej inteligencji
Multivariate anomaly
Statistically abnormal observation detected by simultaneously analyzing relationships and interactions between multiple variables, where each variable individually may appear normal.
Variational autoencoder
Probabilistic neural network that learns the latent distribution of normal data and detects anomalies by their low probability under this learned distribution.
Robust covariance matrix
Statistical estimator resistant to extreme values, calculating relationships between variables while minimizing the influence of anomalies on the estimation.
Hotelling's T-squared
Multivariate statistic generalizing the t-student test, measuring the distance of an observation to the distribution center while accounting for the covariance structure.
Gaussian Mixture Model (GMM)
Probabilistic model representing data as a combination of multiple Gaussian distributions, where low probabilities indicate abnormal observations.
Multivariate DBSCAN
Density-based clustering algorithm identifying anomalies as points not belonging to any dense cluster in multidimensional space.
Isolation Map
Multidimensional visualization technique preserving local relationships to help visually identify clusters and anomalies in complex data.
Robust PCA
Variant of principal component analysis resistant to anomalies, separating data into low-rank subspace and sparse matrix of anomalies.
Copula-based detection
Method that models the dependency structure between variables independently of their marginal distributions to detect multivariate anomalies.
Minimum Covariance Determinant (MCD)
Robust estimator that finds the subset of observations with the minimum covariance determinant, providing an estimate resistant to extreme values.