Implementation Big Data Analysis on Football Competitions via K-means Based Tactical and Generalized Linear Model
Download as PDF
DOI: 10.23977/etemss.2018.1671
Author(s)
Xin Xiang, Guowen Qi
Corresponding Author
Xin Xiang
ABSTRACT
Big data analysis has been applied for technical and tactical performance evaluation in football industry and is prevalent in a decade. It is apparent that novel techniques such as mobile, e-commerce and big data are integrated in this area to inspire new innovations. This paper assesses the technical and tactical performance indicators and the competition results of 2017 China Football Association Super League Tournament via K-means cluster analysis. Competition scores are divided into two categories, balanced and non-balanced scores. A generalized linear model is built for each technical and tactical performance indicators. The competition results in score-balanced matches are defined by linear model to decipher the correlations between the performance indicators and the win probability of the competition. Furthermore, magnitude-based inferences is adopted to define the significance of the linear relationship between tactical performance indicator and the win chance of the game. The developed big data based on K-means algorithm with linear model is available to evaluate the football match performance and facilitate to design task-oriented training plan.
KEYWORDS
K-means algorithm, Tactical, Performance indicator