###What Predicting next matchweek's 10 matches their outcome given in probablities per home, draw and away result. ###How Automatic updating of past match details, including scraping lineups. Taking the next matchweek matches and combine this with expected lineups. Based on the past match details a eXtreme Gradient Boosted tree (XGBoost package) is trained and used to predict next matchweek matches.
Continously
- Historic match details: http://www.football-data.co.uk
- Historic line ups: https://www.11v11.com
- Expected line ups: https://www.rotowire.com
(Half) Seasonly
- Market Values: https://www.transfermarkt.co.uk
- Distance between venues: http://www.sportmapworld.com
Sqlite database
ENG_matches_hist- Match details of historic matches are saved in this table
ENG_match_odds- Match odds of historic matches (odds from InterWetten, LB and William Hill on Home, Draw and Away)
ENG_match_lineup- Historic starting eleven line ups
ENG_players_mv- Market Value of (most) players in the Premier League from 2006 onwards
ENG_exp_lineups- Expected starting eleven lineups for the next matchweek games
ENG_R_lineup_input- Transfer table for importing historic starting eleven line ups before adding them to
ENG_match_lineup. Importing to this table is done throughENG_db_updating.Rand exporting toENG_match_lineupis through the triggerR_lineup_update
- Transfer table for importing historic starting eleven line ups before adding them to
2017