A User-based Collaborative Filtering System for Deck Recommendation in Game Clash Royale
Linfeng Zang, Wenyi Luo · 2022
The decision-making process is often crucial in video games. However, this process may be hard for players to master since it often requires players to accurately analyze and balance various factors. To assist players in making choices, several machine learning algorithms were proposed, one of which is Collaborative Filtering. In this paper, a Collaborating Filtering model improved by user profiles was developed based on the context of Clash Royale, a strategic game. The collaborative Filtering model served to help decide players' decks, collections of cards necessary to battles. Its performance for different players database was evaluated, which showed reliable predicted winning rate that owns errors less than 10% with real testing rate. The evaluation of the model shows that the model can achieve a performance equal to the skilled player, for the winning rates of recommended decks are similar to that of decks used by skilled players.