Classification of top male tennis players

The main objective of this study was to define different quality groups of tennis players based on their position on the ATP ranking list. Ranking data on the top 300 players from 1990 to 2008 were used to conduct the study. The classification into quality groups was performed using six machine learning algorithms suiting such a task. Quality groups were formed better for each year separately than for all years together. Three clustering algorithms (k-means with a Euclidean metric, MDBC with a Euclidean metric, and Xmeans) were equally successful in the classification according to the criteria function. All three algorithms also created very similar quality groups. They are thus equally suitable for classifying tennis players into quality groups based on their ranking scores. Changes in the ranking system (in the year 2000) were also reflected in the differences in classification success between the two periods (before and after 2000). The boundaries between the quality groups were more stable for the period after 2000, and less stable for the period before 2000.
© Copyright 2014 International Journal of Computer Science in Sport. Sciendo. Kaikki oikeudet pidätetään.

Aiheet: tennis huippu-urheilu miespuolinen järjestys matemaattis-looginen malli ryhmä
Aihealueet: tekniset ja luonnontieteet urheilukilpailut
Tagging: Ranking
Julkaisussa: International Journal of Computer Science in Sport
Julkaistu: 2014
Vuosikerta: 13
Numero: 1
Sivuja: 36-42
Julkaisutyypit: artikkeli
Kieli: englanti (kieli)
Taso: kehittynyt