E-sports is an industry with a huge base and the number of people who pay attention to it continues to rise. The research results of E-sports prediction play an important role in many aspects. In the past game prediction algorithms, there are mainly three kinds: neural network algorithm, AdaBoost algorithm based on Naïve Bayesian (NB) classifier and decision tree algorithm. These three algorithms have their own advantages and disadvantages, but they cannot predict the match ranking in real time. Therefore, we propose a real-time prediction algorithm based on random forest model. This method is divided into two stages. In the first stage, the weights are trained to obtain the optimal model for the second stage. In the second stage, each influencing factor in the data set is corresponded to and transformed with the data items in the competition log. The accuracy of the prediction results and its change trend with time are observed. Finally, the model is evaluated. The results show that the accuracy of real-time prediction reaches 92.29%, which makes up for the shortage of real-time in traditional prediction algorithm.
Liu, Jiaxin; Huang, Jiaxin; Chen, Ruyun; Liu, Tianchang; and Zhou, Liang, "A Two-Stage Real-time Prediction Method for Multiplayer Shooting E-Sports" (2020). ICEB 2020 Proceedings. 46.