PCA initialization about minisom

[JustGlowing]

@racng see here racng@618eb5e#commitcomment-139246212

[JustGlowing]

I'll look into this asap, in the meanwhile please refer to this conversation #147

[emeeubc]

Thank you very much for looking into this. I have read conversation #147 and that change to indexing with pc[:,pc_order[0]] makes sense for column-wise eigenvectors. In the snippet of code above, I use pc[pc_order[0]] for indexing because the principal components (pc) are row-wise.

[JustGlowing]

Thanks for this @emeeubc I had a look at you suggestion and your snippet of code is correct. Would you be happy to raise a pull request with the suggested changes?

[emeeubc]

Yes no problem, I've made the pull request. Please let me know if there are any issues. Thanks for your quick response!

[JustGlowing]

issue resolved

[racng]

Can there be an option to use the previous PCA calculation? I have data where the number of rows is much greater than the number of columns. The new PCA calculation is very slow for my data since dimension of cov(data) is very large.

Or would it be wrong to calculate PCA using the previous method?