Cancer type classification experiments: train pan-cancer without X, test X

greenelab / pancancer-evaluation

Evaluating genome-wide prediction of driver mutations using pan-cancer data

BSD 3-Clause "New" or "Revised" License

9 stars 3 forks source link

Closed jjc2718 closed 3 years ago

jjc2718 commented 3 years ago

This should in almost every case be worse than train X, test X - will be a good way to verify that our assumptions about this problem are correct.

jjc2718 commented 3 years ago

Addressed by #30