I am working on 3D data and am using the PCA but I get non-linear effects in sizing and translation when plot the modes of variation.
I don't completely understand the question, so here is just some example matlab code to generate some random data, compute the pca vectors, and then plot them along with the original data.
x = rand(100,3); % This creates 100 random data points.
v = pca(x); % Compute PCA.
scatter3(x(:,1),x(:,2),x(:,3)); %Plot the data.
hold on % Allows me to add to the existing plot, rather than overwrite it.
axis image % This corrects problems with the aspect ratio of the axes.
for i = 1:length(v)
plot3([0,v(1,i)],[0,v(2,i)],[0,v(3,i)]); % Plot a line in the direction of the current pca
% component.
end
- Badges
- Science topic