We developed multiple machine learning models trained on a dataset sourced from Kaggle, which contains data on goals scored by various players in FIFA matches, along with information about the teams they played against. These models were saved and then loaded into the main.py script. The script accepts inputs such as the player's name, the player's team, the opponent team, and whether the player's team won or lost (denoted by is_winner as 0 or 1). Based on these inputs, the model predicts the number of goals the player is likely to score. The script is designed to take these arguments so that it can be executed on a Node.js server from our mern app.

In addition to predicting the number of goals, the script also provides an explanation for the prediction. This explanation is based on SHAP (SHapley Additive exPlanations) and statistical analysis conducted on the dataset, giving insights into the factors influencing the predicted outcome.