Skip to content

Latest commit

 

History

History
24 lines (19 loc) · 1.58 KB

README.md

File metadata and controls

24 lines (19 loc) · 1.58 KB

march-madness-2018

This project uses the data provided by the Kaggle Machine Learning Mania challenge and generates predictions for March Madness brackets. More about the competition: https://www.kaggle.com/c/mens-machine-learning-competition-2018

Feature Engineering

  • power_5: used indicator directly from TeamConferences.csv
  • preseason_rank: found the first ranking using minimum day of ranking from MasseyOrdinals.csv
  • champ: used ConferenceTourneyGames.csv to find the winning team for the last day of conference play for given season, then converted to binary for if the team was the champ that season
    • validated with ConferenceChamps_2018.csv using data from cbssports.com

References:

Feature generation and selection:
Elo scores: