NBA ANALYTICS: Predicting Player Performance via Machine Learning

Postgraduate Thesis uoadl:3243449 66 Read counter

Unit:
Speciality Business Administration, Analytics and Information Systems
Library of the Faculty of Economics and of the Faculty of Business Administration
Deposit date:
2022-11-14
Year:
2022
Author:
Manis Dimosthenis
Supervisors info:
Βασίλειος Λαζάρου, Λέκτορας ,Τμήμα Οικονομικών Επιστημών, Εθνικό και Καποδιστριακό Πανεπιστήμιο Αθηνών.
Original Title:
NBA ANALYTICS : Πρόβλεψη με την βοήθεια μηχανικής μάθησης
Languages:
Greek
Translated title:
NBA ANALYTICS: Predicting Player Performance via Machine Learning
Summary:
In this paper the importance of data analysis and machine learning usage from NBA teams is presented. In the first section, a SWOT analysis of NBA teams that adopted data analysis and machine learning techniques is shown. Strengths and weaknesses will include the direct advantages and disadvantages from analytics adoption and opportunities and threats will include future positive and negative effects that may occur. After SWOT analysis an introduction is made, about machine learning types and some basic algorithms. Next, the thesis goal and thesis methodology are defined, which is predicting a player career performance from stats of its first rookie years. The performance measured with EFF career stat. First, we prepared the data and made an explanatory analysis on them. Then we apply machine learning algorithms in our dataset. Specifically linear regression, decision tree and gradient tree boosting are used. Subsequently adjustments in algorithms have been made, based on problems that occurred. Finally results and conclusions are presented.
Main subject category:
Technology - Computer science
Keywords:
NBA Analytics, Machine Learning, Linear Regression, Decision Trees
Index:
No
Number of index pages:
0
Contains images:
Yes
Number of references:
37
Number of pages:
71
ΕΡΓΑΣΙΑ.pdf (3 MB) Open in new window