From reading paper of Clark-Murphy, 2010, I saw the authors used two methods: "cluster analysis"(CA) and "discriminant analysis" (DA)

I found a great source of explanation here. Essentially speaking, I understand the cluster analysis quite clear that:

Cluster analysis is somewhat exploratory. It takes a data set and looks for the "best" cluster solution or grouping of the people based on their data. Essentially you are trying to find the "true" grouping pattern. Cluster analysis tries to maximize in-group homogeneity and maximize between-group heterogeneity.

However, in this link, I found a lot of discussion about discriminant analysis but I still not yet got the idea. Is there any intuitive explanation for that?

Thanks in advance.

  • 1
    $\begingroup$ Hi @Louise. Overall, cluster analysis (CA) and linear discriminant analysis (LDA) are dimensionality reduction methods. CA methods such as k-means and k-medoids are considered a type of 'unsupervised learning.' Here, these methods are unsupervised because they learn to classify data unconditional of a dependent variable. LDA by comparison is a type of supervised learning, which classifies subsets of data conditional upon a dependent variable. Alongside this, you will find that these methods are typically differentiated by the type of data you have available (whether continuous or categorical) $\endgroup$
    – EB3112
    Nov 8, 2021 at 8:44
  • $\begingroup$ @EB3112 a geat explanation, thanks a heap, it is more clear to me now. $\endgroup$ Nov 8, 2021 at 9:25


Your Answer

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

Browse other questions tagged or ask your own question.