TADPole(data, window.size = NULL, k = 2, dc, error.check = TRUE)cl: Cluster indices.centers: Indices of the centers.distCalcPercentage: Percentage of distance calculations that were actually performed.dtwclust.TADPole clustering adopts a relatively new clustering framework and adapts it to time series clustering with DTW. See the cited article for the details of the algorithm.
Because of the way the algorithm works, it can be considered a kind of Partitioning Around Medoids (PAM). This means that the cluster centers are always elements of the data.
The algorithm first uses the DTW's upper and lower bounds to find series with many close neighbors (in
DTW space). Anything below the cutoff distance (dc) is considered a neighbor. Aided with this
information, the algorithm then tries to prune as many DTW calculations as possible in order to accelerate
the clustering procedure. The series that lie in dense areas (i.e. that have lots of neighbors) are taken
as cluster centers.
The algorithm relies on the DTW bounds, which are only defined for time series of equal lengths.
The windowing constraint uses a centered window. The calculations expect a value in window.size
that represents the distance between the point considered and one of the edges of the window. Therefore,
if, for example, window.size = 10, the warping for an observation $x_i$ considers the points
between $x_{i-10}$ and $x_{i+10}$, resulting in 10*2 + 1 = 21 observations falling within
the window.