Will Townes
Will Townes
Note to self, it would be good to check for this condition and drop the extra column automatically in the initialization code.
Yikes, still pretty weird! What happens if you set L=2 and/or set the learning rate to something small like .0001 ? To do the latter, you can pass it in...
Ohhh, that makes more sense that they are from tSNE. Please try modifying the perplexity parameter on the tSNE to see if that makes a difference. It might also be...
Sorry, I haven't had a chance to work on this, and probably won't for some time. If anyone else gets a chance to figure it out I'd be curious to...
Yes, there is an option called `init` where you can pass in pre-estimated factors, loadings, or both. I have tried it a few times but usually it takes just as...
Thanks for the suggestion, we ended up implementing the feature selection and residuals approximations in a separate package: https://www.bioconductor.org/packages/release/bioc/html/scry.html I will leave this open as a reminder to myself to...
Hi, it's a good question and I honestly don't know whether it's a good idea or not! When I wrote the GLM-PCA paper I didn't think about data integration at...
Need to look at `ilfunc`
Tagging @liuzhe93 so she can keep track of progress here.
Yes, absolutely that is possible in principle, however it's not currently implemented. One difference between this implementation and that of Collins is we do not require the factorized parameters to...