Betting is illegal in India but skilled betting apps like dream11 and betway are currently legal because there’s skilled required to win rather than just the chance or luck. Premier League is a football league which is the biggest sporting league in the world. Machine learning is a subset of artificial intelligence (AI) in which algorithms learn by example from historical data to predict outcomes and uncover patterns that are not easily spotted by humans. Machine learning evolved from the study of pattern recognition and explores the notion that algorithms can learn from and make predictions on data. And, as they begin to become more ‘intelligent’, these algorithms can overcome program instructions to make highly accurate, data-driven decisions. Predictive analytics encompasses a variety of statistical techniques (including machine learning, predictive modelling and data mining) and uses statistics (both historical and current) to estimate, or ‘predict’, future outcomes. These outcomes might be behaviors of a customer likely to exhibit or possible changes in the market, for example. Predictive analytics help us to understand possible future occurrences by analyzing the past. In this project it’s intended to combine machine learning algorithm with predictive analytics to predict the premier league football table at the end of the season.
W: Win
D: Draw
L: Loss
GF: Goals For (Goals Scored)
GA: Goals Against (Goals Conceded)
GD: Goal Difference
Xg: Expected Goals
Xg GD: Expected Goal Difference
xPTS: Expected Points
CLS: Clean Sheets
Bottom 3 Teams Got Relegated so we won't be using their Data as they won't be in the league this season
The result of this project which has the metrics score of 86%. This model is very different when compared to models that are traditionally used by betting companies. It’s fast and easy to understand. When only training 45% of one season’s dataset, it is 40% accurate which is very decent in sports prediction. On the other hand in this project the features which are related to the players are not included so the confidence is not measured and the accuracy is lower. Machine learning can be used to predict in different field, including sports, especially football. Our model is good enough to determine betting odds but obviously it’s not extremely accurate to predict the exact table and exact points scored by each team.
It would be better to add more datasets features like corner goals scored, penalties conceded and do a comparative work to get a more accurate and confident model. By adding some more features like transfer market activities of each club and in which area (defense, offense, midfield). Analyzing more to get which team needs strengthening in which area of the pitch. Testing on more ML algorithms to find out which algorithm will create the most accurate model.