Clustering the trajectories

a.g.leontiev

Hello everybody,

I am trying to learn the new features in mousetrap (mt_cluster). My goal is to see if people who score high on a certain questionnaire have more trajecotories of cluster 1, while those who score low have more trajectories of cluster 2. My question is the following: how can I "mark" each trajectory (each trial), after clustering, as belonging to cluster 1,2 et cetera? So I can then calculate frequencies of trajectories per person. (i.e., subj. A has 5 trials of cluster 1, 6 trials of cluster B, and so on.)

thank you!

Pascal

Hi!

let's start with an example using the example dataset from mousetrap:

library(mousetrap)
mt_example <- mt_spatialize(mt_example)
mt_example <- mt_cluster(mt_example)

By default, mt_cluster stores the cluster of each trial in a new data.frame called "clustering" within mt_example. If you would like to merge it with the trial data, you could simply create a new data.frame that merges them:

results <- merge(mt_example$data, mt_example$clustering, by="mt_id")

Alternatively, you could directly specify that mt_cluster should store the data within mt_example$data via the save_as argument:

mt_example <- mt_cluster(mt_example, save_as="data")

See also the online documentation of mt_cluster for additional information.

Hope this helps!

Best,

Pascal