0
$\begingroup$

I am interested in Hierarchical Clustering Using Mutual Information. Asking the ChatGpt, I got this:

    # Install and load the "mclust" package
install.packages("mclust")
library(mclust)

# Generate some example data
set.seed(1)
data <- matrix(rnorm(100), ncol = 2)

# Perform mutual information clustering using mclust
result <- Mclust(data, G = 2, method = "MICLUST")

# Get the cluster assignment for each data point
clusters <- result$classification

# Print the cluster assignment
print(clusters)

This confused me; there is no "MICLUST" method available in the mclust package, but the code is still working and provides me with a result. Do you know how I can manually use the mutual information for clustering purposes? there is a hclust package in R for hierarchical clustering. Can I modify the similarity function and use a mutual information matrix instead?

Improve this question
$\endgroup$
2
  • 3
    $\begingroup$ It might be that mclust uses the default method (because MICLUST does not exist) or that it simply ignores keyword argument method, because the function does not have such an argument. This is not unusual when using ChatGPT - googling for clustering using mutual information in R may produce more reliable results. $\endgroup$
    – Roger V.
    Commented Jul 1 at 9:13
  • $\begingroup$ @RogerV. Thank you very much for your comment. It is so much appreciated. I googled it, but no results. So, I was thinking of modifying the hclust function. But I don't know if that is correct. $\endgroup$
    – Dr. Statistics
    Commented Jul 1 at 9:18

0

Reset to default

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.