A Study on Machine Learning Algorithms Selection for Estimating the Impact of FIFA World Cups

This study aims to identify the most efficient and accurate machine learning models for precisely estimating the impacts of major sporting events (MSEs). To address these complex challenges, this study broadens the horizons of machine learning applications by integrating sophisticated selection techniques including Stepwise Regression, Support Vector Machines (SVMs), Boosted Neural Networks, Generalized Regression Lasso, Bootstrap Forest, and K-Nearest Neighbors (KNN). This study highlights the diverse nature of variables, including long-term investments, visitor numbers, GDP, and sporting factors such as team scores and performance, using various machine learning models to identify the most suitable models for such datasets. It assessed the contribution of predictors and the significance of variables during the model training and validation phases. Results revealed that the stepwise regression method exhibited robust performance during the training phase. The neural-boosted model exhibited limitations in generalization, demonstrating the intricate nature of and obstacles associated with forecasting the diverse consequences of mega sports events.