Prediction of English Premier League Game Results by Using Deep Learning Techniques  
Author Jaehyun Yi


Co-Author(s) Soowon Lee


Abstract Predicting the outcome of a sporting event can provide an overall flow of events and enable the team to improve its performance. Therefore, much research has been conducted on the prediction of the outcome of sports events through statistical techniques and deep learning techniques. In this paper, we propose a deep learning model predicting English Premier League game results based on the league rankings at the time of the competition and the previous game content of both teams. The proposed model uses the long short term memory model, considering the characteristics of the time series input data. Experimental results show that the average performance of proposed model was 3.6% higher than the existing models.


Keywords Football, deep learning, predictions, Time series data, Logistic regression, LSTM
    Article #:  DSBFI19-96
Proceedings of ISSAT International Conference on Data Science in Business, Finance and Industry
July 3-5, 2019 - Da Nang, Vietnam