ABSTRACT
With the wide spreading of network and capital inflows, Electronic Sport (ES) is developing rapidly in recent years and has become a competitive sport that cannot be ignored. Compared with traditional sports, the data of this industry is large in size and has the characteristics of easy-accessing and normalization. Based on these, data mining and machine learning methods can be applied to improve players' skills and help players make strategies. In this paper, a new approach predicting the outcome of an electronic sport DOTA2 was proposed. In earlier studies, the heroes' draft of a team was represented by unit vectors or its evolution, so the complex interactions among heroes were not captured. In our approach, the outcome prediction was performed in two steps. In the first step, Heroes in DOTA2 were quantified from 17 aspects in a more accurate way. In the second step, we proposed a new method to represent a heroes' draft. A priority table of 113 heroes was created based on the prior knowledge to support this method. The evaluation indexes of several machine learning methods on this task have been compared and analyzed in this paper. Experimental results demonstrate that our method was more effective and accurate than previous methods.
- Agarwala, A., Pearce, M. 2014. Learning Dota 2 team compositions. Technical report, Stanford University.Google Scholar
- Conley, K., Perry, D. 2013. How does he saw me? A recommendation engine for picking heroes in Dota 2. Technical report.Google Scholar
- Kalyanaraman, K. 2014. To win or not to win? A prediction model to determine the outcome of a DotA2 match. Technical report, University of California San Diego.Google Scholar
- Semenov, A., Romov, P., Korolev, S., Yashkov, D., & Neklyudov, K. 2016. Performance of Machine Learning Algorithms in Predicting Game Outcome from Drafts in Dota 2. International Conference on Analysis of Images, Social Networks and Texts (pp.26--37). Springer, Cham.Google Scholar
- Wang, K., & Shang, W. 2017. Outcome prediction of DOTA2 based on Naïve Bayes classifier. Ieee/acis, International Conference on Computer and Information Science (pp.591--593). IEEE.Google Scholar
- Johansson, F., & Wikström, J. 2015. Result prediction by mining replays in dota 2.Google Scholar
- Yang, Y., Qin, T., & Lei, Y. H. 2016. Real-time esports match result prediction.Google Scholar
- Wijaya, A. D., Baskara, F., Kusuma, M. R., Bernhard, R. S., & Senjaya, R. 2016. Ease of use analysis of in-game interface design in DotA 2 for beginners using SAW method. International Conference on Information & Communication Technology and Systems (pp.123--126). IEEE.Google Scholar
- Drachen, A., Yancey, M., Maguire, J., Chu, D., Wang, I. Y., & Mahlmann, T., et al. 2015. Skill-based differences in spatio-temporal team behaviour in defence of the Ancients 2 (DotA 2). Games Media Entertainment (pp.1--8). IEEE.Google Scholar
- Pedregosa, F., Gramfort, A., Michel, V., Thirion, B., Grisel, O., & Blondel, M., et al. 2013. Scikit-learn: machine learning in python. Journal of Machine Learning Research, 12(10), 2825--2830. Google Scholar
Index Terms
- Outcome prediction of DOTA2 using machine learning methods
Recommendations
Using Machine Learning to Predict Game Outcomes Based on Player-Champion Experience in League of Legends
FDG '21: Proceedings of the 16th International Conference on the Foundations of Digital GamesLeague of Legends (LoL) is the most widely played multiplayer online battle arena (MOBA) game in the world. An important aspect of LoL is competitive ranked play, which utilizes a skill-based matchmaking system to form fair teams. However, players’ ...
An Analysis of DOTA2 Using Game Refinement Measure
Entertainment Computing – ICEC 2017AbstractDotA is one of the most attractive and influential MOBA games, which has been popular in many countries for over 10 years. It was designed for fun at first, however, after the emerge of DOTA2 it soon became the highest prize e-sports game. This ...
Champion Recommendation in League of Legends Using Machine Learning
Computational Science – ICCS 2023AbstractLeague of Legends (LoL) is a Multiplayer Online Battle Arena (MOBA) game with over 160 champions and a competitive esports scene. The ban-and-pick system allows players to choose champions before a match, which can greatly impact the outcome, ...
Comments