This is an old revision of the document!
Dynamic Time Warping
Literatur Iterative Deepening Dynamic Time Warping for Time Series
PDTW
Dynamic Time Warping mit Piecewise Aggregate Approximation (PAA): Approximieren einer Zeitreihe durch Segmentierung in gleichlange Teile und Mittelwert der Datenpunkte innerhalb dieser Punkte.
Multidimensional DTW (MD-DTW)
Zuerst Normalisieren jeder Dimension von t und r.
$$d(i,j) = \sum_{k=1}^K(t(k,i)-r(k,j))^2$$
Literatur
A Scalable Method for Time Series Clustering
Ersetzen der Punkte durch globale charakteristische Measures.
Dynamic Time Warping (DTW) has been applied in time series mining to resolve the difficulty caused when clustering time series of varying lengths in Euclidean space or containing possible out-of-phase similarities (Berndt & Clifford, 1994; Keogh, 2002; Ratanamahatana & Keogh, 2004). […] but is not defined if a single data point is missing.