Retrieve variables from PCA generated by K-means clustering
Posted: Thu Mar 07, 2024 7:16 am
Hi!
I'm trying to cluster a data set containing approximately 100 variables into two clusters using K-means clustering in Jamovi. Under "Plots" I tick Variables-PCA to generate a PCA which I assume indicate how and to what extent my different parameters drive clustering. In the "Variables-PCA" plot, the variabels appear by name, with an arrow and are also colored in accordance with my two clusters. I interpret the PCA as that variables that are assigned as cluster 1 drive grouping of samples into cluster 1 and that the variables with the longest arrows are the ones that have the biggest impact on the cluster (and opposite for variables that are assigned as cluster 2). Is there a way by which I can retrieve this information into the format of a table to be able to identify the variables that to the highest extent contribute into separating the two clusters?
Best,
H
I'm trying to cluster a data set containing approximately 100 variables into two clusters using K-means clustering in Jamovi. Under "Plots" I tick Variables-PCA to generate a PCA which I assume indicate how and to what extent my different parameters drive clustering. In the "Variables-PCA" plot, the variabels appear by name, with an arrow and are also colored in accordance with my two clusters. I interpret the PCA as that variables that are assigned as cluster 1 drive grouping of samples into cluster 1 and that the variables with the longest arrows are the ones that have the biggest impact on the cluster (and opposite for variables that are assigned as cluster 2). Is there a way by which I can retrieve this information into the format of a table to be able to identify the variables that to the highest extent contribute into separating the two clusters?
Best,
H