0
$\begingroup$

I have a mixed data and I have been searching for the best method to cluster it and I've chosen PAM. I am working btw with R. I've considered all the 17 variables of my data to cluster: 4 qualitative and the rest is continuous. When I plot my clustering, I got the following message under the plot: These two components explain 4.43% of the point variability. Which is very little right? Does that mean that method or the quality of my clustering is/are bad? what would be other measures for my clustering quality? What would recommend to me as a clustering method or as a way of improving the results of my PAM clustering?

The thing is that the visualization did work and also the distance function?But that very little point of variability the two components are explaining is source of worry! (My data is of almost 4000 observation with 17 var). I have to say that I am very new to R and data analysis in general!Any concise advise would be helpful for me! Thank you Happy new year!

Thanks a lot for your answer. The thing is that the visualization did work and also the distance function?But that very little point of variability the two components are explaining is source of worry! (My data is of almost 4000 observation with 17 var). I have to say that I am very new to R and data analysis in general!Any concise advise would be helpful for me! Thank you enter image description hereenter image description here

Improve this question
$\endgroup$
1
  • $\begingroup$ Please use the gray edit link under your question to add additional information. $\endgroup$
    – gung - Reinstate Monica
    Commented Jan 1, 2019 at 14:32

1 Answer 1

Reset to default
1
$\begingroup$

It means the two dimensional plot is not very accurate.

So the automatic plotting routine doesn't work well for your data.

You'll need to come up with an appropriate - different - way if visualizing and analyzing your clustering results yourself.

And it's important to do so. Because when the visualization doesn't work, there is a good chance your distance function also didn't work and hence your clustering didn't work...

Bad news for you is that the standard approaches failed... You'll need to carefully find your own way.

Improve this answer
$\endgroup$

Your Answer

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

Not the answer you're looking for? Browse other questions tagged or ask your own question.