abidlabs / contrastive

Contrastive PCA
MIT License
199 stars 47 forks source link

trying to understand how this is unsupervised learning/theory behind #23

Open klai001 opened 8 months ago

klai001 commented 8 months ago

Dear all thank you for developing this method! its very useful indeed. I would like to understand more about this method/ i.e. how does this consider as unsupervised learning when the user defines the background and target datasets e.g. healthy vs disease? and how does it differ from the purpose of e.g. PLSDA? where class labels/categorial covariates are also specified?