Extracting the roles of different players in soccer using an automatic clustering algorithm

چکیده مقاله

Nowadays analyzing team sports using artificial intelligence algorithms has become one of 
the most interesting topics for data scientists. On the other hand, it is really important for the 
coaches and managers to analyze their own and opponents’ performance. So team sport 
companies support data scientists to get good results based on the valuable knowledge 
extracted by the data scientists from massive and unstructured datasets. Clustering, which is 
the process of grouping data points in a dataset in to several clusters based on their 
similarities, is one of the most important data mining and big data mining tools which has 
been widely used in different fields. In this research a new automatic big data clustering 
algorithm, based on a swarm intelligence method, is used to cluster a big dataset containing 
centers of players’ performance in different soccer matches to extract the role of different 
players in a soccer match. The proposed method consists of two phases. In the first phase 
the algorithm tries to find proper number of clusters and in the second phase it tries to find 
the position of the centroids. The dataset which is used in this research contains the center 
of several players’ performance in about 93000 soccer matches which means that the dataset 
has 93000 objects