Pattern Recognition in Games Using Process Mining
[ 1 ] Instytut Informatyki, Wydział Informatyki i Telekomunikacji, Politechnika Poznańska | [ P ] employee
2022
chapter in monograph / paper
english
- Process Mining
- Machine Learning
- Artificial Intelligence
- Gaming
- Pattern Discovery
EN Nowadays, one of the most rapidly growing entertainment markets in the world is the video gaming industry. Modern games put immersion of gameplay into focus providing, appropriate gameplay balance to make the game challenging enough to be interesting and not too hard to make it fun. Therefore, developers spend a huge amount of resources and time to test games, analyze the feedback and improve game balance. Unfortunately, games are becoming more and more complex, which makes them harder to analyze by hand. To solve this issue developers are looking for an efficient and automatic method to find information about an unbalanced element of gameplay. Therefore in this article authors tackle the problem of discovering knowledge about video game balance based on pattern discovery (recognition) of players' behavior. First, the authors discuss how to gather appropriate data. Secondly, they propose a player behavior pattern discovery algorithm (Gameplay Miner, GP). Then, they show how the results of pattern recognition can be visualized and analyzed. As a result, overpowered and underpowered gameplay elements are identified. The paper ends with a performance test that shows how the size of the input data is related to the runtime and memory needed for the algorithm.
18.08.2022
42 - 45
20