Machine learning is a rapidly grown term which has application in various fields of our
life one of them is in sports analytics.
In this paper we made use of data, which were extracted from 19 seasons of NBA
games. The goal of the thesis is to exploit the data we had trying to measure every
team’s game performance and predict their final position after NBA Playoffs.
Extracted data concerns not only the most fundamental team statistical categories but
also some miscellaneous features regarding team performance. In this thesis firstly we
attempt to depict the development of the NBA industry within the years along with the
change of the game’s nature itself. Furthermore, with the analytics tools we can extract
valuable information regarding which are the key factors that can lead a team in the top
of NBA championship.
We conducted several experiments using a variety of classifiers aiming to predict the
number of wins of every team participating in Playoffs. Despite of the new format at
playoffs results were accurate and very interesting assumptions were made.
Collections
Show Collections
