Open Access
American Research Journal of Computer Science and Information Technology
ISSN (Online): 2572-2921
DOI: 10.46568/arjcsit
Optimization of Big Data Processing Algorithms for Cyber Sports Platforms
Abstract
This study aims to optimize big data processing algorithms for esports platforms, addressing the unique challenges
posed by the industry’s rapid growth and data complexity. The research employs a multifaceted approach, combining
adaptive stream processing techniques, multi-level data storage architectures, and ensemble machine learning models.
Key innovations include a dynamic window sizing algorithm for real-time data analysis, a hybrid LSTM-XGBoost model
for player performance prediction, and an adaptive load balancing system based on multi-agent theory. The proposed
framework demonstrates significant improvements in processing latency (40% reduction), system throughput (65%
increase), and prediction accuracy (25% enhancement) compared to traditional methods. These advancements lay the
foundation for next-generation analytical platforms in esports, offering potential applications in other domains requiring
real-time processing of dynamic, large-scale datasets. The study’s novelty lies in its tailored approach to esports-specific
data characteristics and its integration of cutting-edge techniques from diverse fields of computer science.
