Use of Self Organizing Maps in Technique Analysis

Authors Roger Bartlett, Peter Lamb, Anthony Robbins



PDF
Thumbnail PDF

File

DagSemProc.08372.8.pdf
  • Filesize: 0.53 MB
  • 8 pages

Document Identifiers

Author Details

Roger Bartlett
Peter Lamb
Anthony Robbins

Cite AsGet BibTex

Roger Bartlett, Peter Lamb, and Anthony Robbins. Use of Self Organizing Maps in Technique Analysis. In Computer Science in Sport - Mission and Methods. Dagstuhl Seminar Proceedings, Volume 8372, pp. 1-8, Schloss Dagstuhl – Leibniz-Zentrum für Informatik (2008)
https://doi.org/10.4230/DagSemProc.08372.8

Abstract

This study looked at the coordination patterns of four participants performing three different basketball shots from different distances. The shots selected were the three-point shot, the free throw shot and the hook shot; the latter was included to encourage a phase transition between shots. We hypothesised lower variability between the three-point and free throw shots compared to the hook shot. The study uses Self-Organizing Maps (SOM) to expose the non-linearity of the movement and to try to explain more specifically what it is about the coordination patterns that make them different or similar. The SOM proved to draw the researcher's attention to aspects of the movement that were not obvious from a visual analysis of the original movement either viewed from video or as computer animation. A speculative link between the observational learning literature on the importance of the kinematics of distal segments in skill acquisition and the visual information a coach or analyst may rely on for qualitative technique analysis was made. Although making the distinction between the three shooting conditions was meant to be a trivial exercise, in many cases for this dataset the SOM output and the natural inclination of the movement analyst did not agree: the SOM may provide a more objective method for explaining movement patterning.
Keywords
  • Artificial neural networks
  • basketball shooting
  • movement coordination
  • movement variability
  • self-organizing maps.

Metrics

  • Access Statistics
  • Total Accesses (updated on a weekly basis)
    0
    PDF Downloads