search

theislab / destiny

R package for single cell and other data analysis using diffusion maps
https://theislab.github.io/destiny/
GNU General Public License v3.0
76 stars 12 forks source link

Different result of Calling DiffusionMap #33

Closed RuixiangLiu01 closed 4 years ago

RuixiangLiu01 commented 4 years ago

Hello developer,I found that every time when I calling DiffusionMap function, I get the different result about eigenvector.But as I know, these vector was computed from eigen decomposition from a Matrix. So, could someone explain why this would happen?How can I do to solve this problem.I installed destiny with devtools::install_github('theislab/destiny').Thanks! image image

flying-sheep commented 4 years ago

The implementation of Eigen decomposition I use is approximate and nondeterministic (everything else would be unbearably slow). The newest destiny versions should make sure that set.seed(...) works as expected and results in the same decomposition being found, see #5 and yixuan/RSpectra#14

I can’t completely guarantee it though, since 3rd party libraries sometimes don’t respect it.

RuixiangLiu01 commented 4 years ago

The implementation of Eigen decomposition I use is approximate and nondeterministic (everything else would be unbearably slow). The newest destiny versions should make sure that set.seed(...) works as expected and results in the same decomposition being found, see #5 and yixuan/RSpectra#14

I can’t completely guarantee it though, since 3rd party libraries sometimes don’t respect it.

sorry for update late, I remove destiny and reinstall with devtools::install_github('theislab/destiny')， the problem was solved.Thank u so much!

flying-sheep commented 4 years ago

No problem, great that it works for you!