Entering edit mode
5 weeks ago
4732b159
•
0
Hello, I did some PCA analysis on my data using factoextra package a while ago, and of course didn't change anything in the code or in the analysis since then. Today I needed to go through the results and re-open R and re-run the exact same code, however, surprisingly I am getting different PCA plots this time with different PC1, PC2.. percentages and pattern. What might be the reason and how to make it reproducible? Thanks for your comments :)
If you said t-SNE or some other stochastic method I would have believed it, but PCA is reproducible. I will take your word that it is the same code, but something must have changed either in your data or in the software. Have you not updated anything since "a while ago"?
Have you set a seed?
Did you switch computer or virtual environment ?