Author : Marck England P. Bautista 1
Date of Publication :28th September 2023
Abstract: Online gaming, particularly video games, is a popular leisure activity. Matchmaking is crucial in e-sports and online gaming as it directly affects player satisfaction and the longevity of gaming products. A proposed solution to address unequal matchmaking in online gaming is to establish a performance-driven system. This study used the Las Vegas Algorithm (LVA) for player selection and K-Nearest Neighbor (KNN) for categorizing and classifying player performance data. This study proposed a hybrid algorithm for online game matchmaking that combined LVA and KNN. The hybrid approach includes improvements such as data classification, runtime optimization, and increased success probability. The study used a dataset of 80,000 raw data and 32 variables that underwent Mutual Information-Based Feature Selection. The study showed that using LVA and KNN together improved data categorization and classification, and a greater probability of success. However, the hybrid algorithm had a longer runtime compared to the Las Vegas algorithm. The hybrid algorithm necessitates an initial data categorization phase prior to selecting players randomly. The existing algorithm disregards player performance when identifying them. The hybrid algorithm takes longer to execute due to the extra computational steps needed for data categorization, which are not present in the current algorithm. Despite its drawback, the hybrid algorithm can enhance player selections by integrating performance rates into the categorization process.
Reference :