nihalb/JMARS

There are many duplicates in data.json

Opened this issue · 0 comments

Hello!

It seems data.json linked in README.md has 9000 duplicates (so ~8%), which can hurt generalization estimation when using standard CV. Is there maybe a corrected version?