Clustering online game communities through SOM

  • Authors:
  • Lia C. Rodrigues;Clodoaldo A. M. Lima;Pollyana N. Mustaro

  • Affiliations:
  • School of Engineering, Mackenzie Presbyterian University, São Paulo, SP, Brazil;School of Engineering, Mackenzie Presbyterian University, São Paulo, SP, Brazil;School of Engineering, Mackenzie Presbyterian University, São Paulo, SP, Brazil

  • Venue:
  • IJCNN'09 Proceedings of the 2009 international joint conference on Neural Networks
  • Year:
  • 2009

Quantified Score

Hi-index 0.00

Visualization

Abstract

Nowadays, online games have an exponential increase in the market because many people interact for hours in a virtual gaming worlds called the Massive Multiplayer Online Role-Playing Games (MMORPGs). In this kind of environment players maintain relationships and build communities. To study the common characteristics and relationships of the communities formed in those games, it is possible to cluster a player's community. Moreover, player's community structure is common in various real-world networks; methods or algorithms for grouping such communities have attracted great attention in recent years. The analysis of those groups aim to better understand and examine the behaviour of players. In this paper, self-organizing maps were explored to obtain clusters of a player community from the game World of Warcraft (WoW). To improve the efficiency of the clustering methodology masks were applied that considered the player's individual score, player's guild degree (number of connections), and player's class. The results obtained indicate that the proposed methodology can be successfully applied to the clustering online game communities.