论文信息 - Analysis of NBA player positions using ISODATA: a case study

Analysis of NBA player positions using ISODATA: a case study

In this paper, we apply the ISODATA to cluster NBA player positions according to their game statistics in regular season 2005-2006. In the experiment, there were totally 214 players used to be examples. Three phases were processed during the experiment based on three different desired clusters (positions). In the first phase, three positions were classified. In the second phase, five positions were classified. In the third phase, seven positions were classified. We used confusion matrices to calculate the clustering accuracies. The experimental results show that less numbers of the desired clusters get better clustering results.

[1] Bing Sun,et al. Hierarchical speaker identification using speaker clustering , 2003, International Conference on Natural Language Processing and Knowledge Engineering, 2003. Proceedings. 2003.

[2] Rasim Latifovic,et al. A Comparison of Clustering Strategies for Unsupervised Classification , 2000 .

[3] Richard J. Roiger,et al. Data Mining: A Tutorial Based Primer , 2002 .

[4] Che-Chern Lin,et al. A schema to determine basketball defense strategies using a fuzzy expert system , 2006 .

[5] M. Isabel García-Planas,et al. Proceedings of the 6th WSEAS international conference on Applied computer science , 2006 .

[6] Shirley Litty. Coaching Girls' Basketball , 1981 .

[7] David M. Mount,et al. A Fast Implementation of the Isodata Clustering Algorithm , 2007, Int. J. Comput. Geom. Appl..

[8] Hamid Soltanian-Zadeh,et al. Multiparametric ISODATA analysis of embolic stroke and rt-PA intervention in rat , 2004, Journal of the Neurological Sciences.

[9] Julius T. Tou,et al. Pattern Recognition Principles , 1974 .