Closed peterwhiteenjine closed 4 years ago
Yes, the results match other than the split. I've put back some of your old code and kept the new split as well as sorting at the end. Our results match those from Marcos' code.
Then it's not really a bug. Please close the PR
On Sun, May 10, 2020, 11:13 AM peterwhiteenjine notifications@github.com wrote:
Yes, the results match other than the split. I've put back some of your old code and kept the new split as well as sorting at the end. Our results match those from Marcos' code.
— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/enjine-com/mcos/pull/44#issuecomment-626367282, or unsubscribe https://github.com/notifications/unsubscribe-auth/AAXTIXKFQQWTNNEZIEG6TZ3RQ3VFTANCNFSM4M3OZ5CA .
Ok. Will do.
Just to make sure all is clear: our code as is produces different weights than Marcos' code because of the different splits. If I change the splitting to match Marcos', we get the same results as his code.
Ok if you correct for the split, does our algo results match Marcos elsewhere?
On Fri, May 8, 2020, 4:48 AM peterwhiteenjine notifications@github.com wrote: