Clustering Time Series Data using dtwclust: Unable to Access Cluster Labels from Multiple Runs

75 views Asked by At

I am working on a clustering analysis of time series data in R using the dtwclust package, specifically the tsclust function. My goal is to perform multiple runs of the clustering and extract cluster labels from each run.

Here's an example code (from dtwclust documentation):

# Load data
    data(uciCT)

    # Reinterpolate to same length
    series <- reinterpolate(CharTraj, new.length = max(lengths(CharTraj)))

    # Subset for speed
    series <- series[1:20]
    labels <- CharTrajLabels[1:20]

    # Making many repetitions
    pc.l2 <- tsclust(series, k = 4L,
                     distance = "L2", centroid = "pam",
                     seed = 3247, trace = TRUE,
                     control = partitional_control(nrep = 10L))

# Cluster validity indices
sapply(pc.l2, cvi, b = labels)

The code works fine, but I'm having trouble accessing cluster labels from multiple runs. Specifically, I want to extract cluster labels from each run. Modifying the code as follow (setting one single run), extracting cluster labels is not a problem

    # Making many repetitions
    pc.l2 <- tsclust(series, k = 4L,
                     distance = "L2", centroid = "pam",
                     seed = 3247, trace = TRUE,
                     control = partitional_control(nrep = 1L))

    # Extracting cluster labels
    labels <- pc.l2@cluster

This allows me to extract cluster labels, but I need to perform multiple runs, and I'm unsure how to access cluster labels from each run.

Can anyone provide guidance on how to access cluster labels from multiple runs of the tsclust function in the dtwclust package?

0

There are 0 answers