README.md

---
license: mit
task_categories:
- text-classification
language:
- en
tags:
- football
- classification
size_categories:
- 10K<n<100K
---
The ID column links tables in X_train, with Y_train and Y_train_supp. The same holds true for the test data.

Input team data sets comprise the following 3 identifier columns:

ID, LEAGUE and TEAM_NAME (note that LEAGUE and TEAM_NAME are not included in the test data)
The following 25 statistics, which are aggregated by sum, average and standard deviation.

'TEAM_ATTACKS'
'TEAM_BALL_POSSESSION'
'TEAM_BALL_SAFE'
'TEAM_CORNERS'
'TEAM_DANGEROUS_ATTACKS'
'TEAM_FOULS'
'TEAM_GAME_DRAW'
'TEAM_GAME_LOST'
'TEAM_GAME_WON'
'TEAM_GOALS'
'TEAM_INJURIES'
'TEAM_OFFSIDES'
'TEAM_PASSES'
'TEAM_PENALTIES'
'TEAM_REDCARDS'
'TEAM_SAVES'
'TEAM_SHOTS_INSIDEBOX'
'TEAM_SHOTS_OFF_TARGET'
'TEAM_SHOTS_ON_TARGET',
'TEAM_SHOTS_OUTSIDEBOX'
'TEAM_SHOTS_TOTAL'
'TEAM_SUBSTITUTIONS'
'TEAM_SUCCESSFUL_PASSES'
'TEAM_SUCCESSFUL_PASSES_PERCENTAGE'
'TEAM_YELLOWCARDS'
Input player data sets comprise the following 3 identifier columns:

ID, LEAGUE and TEAM_NAME, POSITION and PLAYER_NAME (note that LEAGUE, TEAM_NAME, and PLAYER_NAME are not included in the test data)
52 statistics, which are aggregated by sum, average and standard deviation. They are similar to the team statistics though more fine-grained.