<![CDATA[As the digital age progresses, the more technology affects various aspects in our lives for example entertainment through video games. A problem arises where there are too many video games to choose from, so there is a need to find methods to narrow down the choices. This study implements a collaborative filtering-based video game recommendation system to analyze user preferences based on playtime data. The system processes user-game interaction data from a secondary dataset containing 14.3 million players and 50.9 million games, constructing a sparse matrix to map user playtime behavior. By identifying similar users through kNN, the system recommends games frequently played by users with comparable preferences. Evaluation on 100 sample users achieved an impressive mean precision of 88.12%, indicating that most recommended games were among the users' top 20 most-played titles. This study hopes to further enable people in finding more fun experiences in their lives.]]>
Copyrights © 2026
