|
FPL-Team-Prediction
This project uses machine learning to predict the fantasy premier league performance of each player.
Data is gathered from various sources such as understat, fpl webste using APIs, etc. The data is prepared and processed using Python libraries.
Three supervised learning models (Linear Regression, Random Forrest and XGBoost) are built and trained for making predictions.
For each gameweek, the models are trained using all historical data prior to that week starting from the 2016/17 season using features such as -
player performance in fpl, player understat statistics, team's form, etc. After points forecast, Linear Programming is used to find an optimal team.
|