Predicting the Football Players’ Market Value Using Neural Network Model: A Data-Driven Approach

The domain of this research is about football business in which the market value of football players was predicted. The researcher designed a neural network model to predict the price of a football player for football clubs utilizing data from FIFA 19. The researcher thoroughly conducted an in-depth research on literature review to understand the inner workings of the neural networks as well as found related works on sports analytics and prediction. On the other hand, this research involved the hyper parameter optimization of the neural network with different activation functions, number neuron and layers, learning rate and its decay and L1 regularization. K-Fold cross validation is carried out to ensure the quality of the neural network. Simultaneous exploration is done to various aspects of neural network training is performed and their evaluation metrics such as mean squared error, R-squared and Root Mean Square are investigated. Lastly, the best performing model is chosen for the final model. The final model achieved a 95 percent accuracy and a 0.037 mean square error.