el-iot / oracle Goto Github PK

FOOTBALL MACHINE LEARNING ENGINE

Oracle was created by Rory

Click here to visit Oracle.

Oracle uses a form of machine learning known as poisson regression to predict the outcome of upcoming Premier League football games based upon past event data.

Bookmaker data adjusts in real time as bets are placed in order to hedge the bookmaker's position on either side of the event. Popular bets where the published odds have become skewed can be identified and ranked by significance using machine learning.

Oracle hunts the internet for bookmaker data in the 3 major betting regions (USA, UK and Australia). It compares the current published odds data with the outcome of the Oracle machine learning engine and suggests the bets with the highest expected value. The strongest bets are those with the highest integer values in the significance column and the most stars in the strength column.

Oracle also provides a more straightforward head to head demonstration mode with win/draw probabilities updated in real time.

Oracle is a demonstration of machine learning only and Oracle assumes no responsibility for the outcome of any games or bets made as a result of this engine. Oracle was created by Rory. Poisson regression for this style of problem was first suggested in papers by David Sheehan and Erlandson F. Saraiva.

View the website here.

Data updated weekly.

Head to head example shown below: