Recommendation of Videogames with Fuzzy Logic

Abstract

A videogame as software and as a product presents a great variety of characteristics: gender, theme, platform, target audience, among others. In recent years, the number of videogames developed has grown notably thanks to the industry, as users have a large catalog available, who often may be curious to play another videogame that has not been presented in an advertising medium. In present investigation we propose a videogame recommendation architecture and a recommendation system using Fuzzy Logic, in the construction we have designed 16 rules with fuzzy sets. A database of approximately 55,000 games and of its 16 attributes of which 5 were used for the recommendation system was used: 4 attributes (Critic Score, User Score, Global_Sales, Year) to establish membership functions with the 16 rules of recommendation formed based on the opinion of experts in the field of videogame analysis and 1 attribute (Age) to develop the content filter according to age applying an ethics model in Artificial intelligence. The results of our computational experiments with the proposed architecture reached an accuracy percentage of 80,0%.