sangitaghosh / analyticsvidhya_xtrememlhack

Winning solution for Xtreme ML Hack on AnalyticsVidhya

Xtreme ML Hack on AnalyticsVidhya

The hackathon was hosted on 20th - 23rd April, 2017. I modelled Contacts and Resolution independently, and it was one of the few times where using simple averages beat all ML-models I tried, so ended up sticking with a weighted average solution.

1. Please edit the paths in the files before running them.
2. I blended this model with Mark's ML-based solution, which gave our best score (64.94) on public LB and ranked 1st on private LB.

Thanks to Mark Landry (Link to his model), for teaming up and sharing some interesting insights and ideas. Hope to collaborate again in the times to come!