Predicting Football Shot Outcomes in the Turkish Super League: Support Vector Machine and Naive Bayes Models with Feature Importance Analysis

This study presents a modeling framework that formulates goal prediction as a binary classification problem (1 for a goal, 0 for no goal), based on shot records from the Turkish Super League between the 2020–2025 seasons. In this study, Naive Bayes (NB) and Support Vector Machine (SVM), which are supervised machine learning algorithms, used for prediction models. The focus of this study is to build a model and improve its overall performance by addressing the significant class imbalance between goals and non-goals in the original dataset. The results of machine learning models created for a balanced dataset were compared, and the findings showed that SVM models were relatively more successful in adapting to the complex data structure of football. Feature importance analyses confirmed that distance is the most dominant predictor of shot success. It has appeared that the most important variables are related to distance and angle. This study contributes to the literature with its predictive power and model transparency, and offers a decision support framework, especially for the Turkish Super League.