Ανάλυση δεδομένων καλαθοσφαίρισης για την πρόβλεψη αποτελεσμάτων με επιλογή χαρακτηριστικών

Μιχαλάκης, Ζαχαρίας

Master Thesis

Author

Μιχαλάκης, Ζαχαρίας

Date

2024-05

Abstract

This Master's thesis delves into the field of Sports Data Analysis from the sport of Basketball in order to predict results and select characteristics that contribute to the achievement of this goal. The paper develops and applies modern machine learning and statistical analysis methodologies to process and evaluate data, seeking to identify which data and statistics have the greatest predictive value. The research begins with an extensive review of the relevant literature, examining previous studies that have dealt with outcome prediction in sport and basketball in particular. The methodology adopted for data processing is then presented, which includes pre-processing, feature selection, and modelling techniques. Special emphasis is given to the application of machine learning algorithms, such as Logistic Regression, k-nearest neighbors (KNN), Support Vector Machine (SVM), Random Forest, Neural Network(Multi-layer Perceptron), to evaluate their predictive ability with respect to game outcomes. Another goal was to make a forecast of the last 5 years with the above classifiers, based on previous years for each dataset. In the following, the results of the application of these methods to a dataset that includes statistics from professional basketball games, specifically the European League (Euroleague), are presented. The analysis focuses on evaluating the importance of each attribute and its contribution to predicting the outcome of the games, as well as the effectiveness of the different prediction models. The ultimate goal was to make outcome prediction with the above categorizers based on the available dataset. Finally, the paper concludes with a discussion of the challenges and prospects of outcome prediction in basketball, suggesting extensions for future research efforts. In addition, the importance of incorporating more data and improving machine learning algorithms to increase the accuracy of predictions is acknowledged. This work contributes to the field of analytical sports science by providing valuable insights for the application of machine learning to sports outcome prediction.

Postgraduate Studies Programme

Πληροφοριακά Συστήματα και Υπηρεσίες

Department

Σχολή Τεχνολογιών Πληροφορικής και Επικοινωνιών. Τμήμα Ψηφιακών Συστημάτων

Number of pages

Language

Greek

URI

https://dione.lib.unipi.gr/xmlui/handle/unipi/16461
http://dx.doi.org/10.26267/unipi_dione/3883

Collections

Τμήμα Ψηφιακών Συστημάτων

Show full item record

Except where otherwise noted, this item's license is described as
Αναφορά Δημιουργού - Μη Εμπορική Χρήση - Παρόμοια Διανομή 3.0 Ελλάδα