KI-Glossar
Das vollständige Wörterbuch der Künstlichen Intelligenz
ARIMA
AutoRegressive Integrated Moving Average statistical model used to analyze and forecast time series by capturing trends, seasonality and autocorrelations in the data.
STL
Seasonal and Trend using Loess decomposition algorithm that separates a time series into trend, seasonality and residual components to facilitate anomaly detection.
CUSUM
Change point detection algorithm that accumulates deviations from a reference mean, signaling anomalies when the cumulative sum exceeds a threshold.
Prophet
Forecasting library developed by Facebook that decomposes time series into trend, seasonality and holidays, facilitating the identification of residual anomalies.
DBSCAN
Density-based clustering algorithm that identifies anomalies as isolated points in low-density regions of the temporal space.
Holt-Winters
Triple exponential smoothing method that models level, trend and seasonality, allowing anomaly detection through forecast error analysis.
LSTM Autoencoder
Hybrid architecture combining LSTM and autoencoder to learn normal temporal representations and reconstruct sequences, with high errors indicating anomalies.