MATLAB Answers


Help with PCA analysis

Asked by Elise
on 19 Sep 2013

I have a set of measurements of wrist movements. It contains data about the movements in x,y,z translations and x,y,z rotations - so it has 6 dimensions.

I used:

[coeff,score,~,~,explained,~] = pca(E); 

Explained gives me the following result:


From this I see that 80% of the data is in the first two principle components.

I assume that these 2 principle components form a plane. How do I use coeff to see how this plane is orientated?

Thanks in advance for helping me out!


the cyclist
on 19 Sep 2013

When you say "see how this plane is orientated", do you literally mean a visualization? It's a challenge to show a (2-D) plane embedded in a 6-D space.

on 20 Sep 2013

Good question, visualizing would be preferable, but I can see that that would be difficult.

I quess what I want to know is how much of each original component can be found in the first two principal components. For this case: how much translation v.s. rotation can be found in the first to principle components.

I know this should be found in coeff/score, I just can not figure out how to interpret these.



No products are associated with this question.

0 Answers

Discover what MATLAB® can do for your career.

Opportunities for recent engineering grads.

Apply today