Machine Learning Applications in Multiplayer Online Battle Arena Esports—A Systematic Review

Machine learning (ML) is an emerging field, while multiplayer online battle arena (MOBA) esports has seen a rise as a research subject of interest. The applications of ML in MOBA are an interesting field of study as the esports genre is enriched by a plethora of data. Moreover, the MOBA esports industry is a budding economic sector with stakeholders looking for innovative scientific studies. This necessitates the need for a systematic review to provide insights for future studies. The databases (Scopus, Web of Science, PubMed, ScienceDirect and Google Scholar) were systematically searched to identify published peer-reviewed academic articles. The Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) method was used for the analysis. Papers were included if they contained ML applications for MOBA, excluding game design or non-esports-related studies. There are 35 studies included in this systematic review, with most studies on Defence of the Ancients 2 (DotA 2) and League of Legends (LoL). ML algorithms are used to make predictions for different purposes using game mechanics and player profiles as datasets, with random forests, decision trees, logistic regression, neural networks, and more. The performance of ML models was deemed impressive, given their simplicity and the datasets used. Key findings highlight the potential future area of research in the empowerment of mobile phone-based MOBA, commercialisation opportunities using ML technology in authentic settings, and overcoming challenges in data access and regional differences.