I have just read about the usual mistake in Principal Components Analysis of confusing between principal directions and component loadings following the explanation and links here:

Relationship between SVD and PCA. How to use SVD to perform PCA?

After clarification of how to extract the loadings when performing PCA dimensionality reduction I was wondering whether once we have extracted the loadings of the components (how important each feature is on each principal component) the simple structure criteria from Thrustone can be used to assess how close we are to an ideal simple structure.

  • 1
    $\begingroup$ I believe thread stats.stackexchange.com/q/612/3277 neatly answers it. $\endgroup$ – ttnphns Jan 17 '17 at 14:10
  • $\begingroup$ Hi, thanks for the answer. I had read the thread before but I still have the doubt that I see in many explanations that PCA ≠ FA but the loadings extracted after performing PCA can resemble the loadings in FA. I understand PCA and FA are different as their statistical formulation differs hence my intuition tells me not to extrapolate the criteria from Thrustone to PCA loadings. Thanks a lot for your help clarifying it. $\endgroup$ – Lumbreras Jan 22 '17 at 22:35

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.