Creative Commons source reading · Prediction and forecasting
Predicting sport event outcomes using deep learning
Jianxiong Gao; Yi Cheng; Jianwei Gao
Originally published in PeerJ Comput Sci · 2025-07-21
From the article
Selected source text. JournalOS removes markup and shortens excerpts where indicated. The full article provides the complete context.
Abstract
Predicting the outcomes of sports events is inherently difficult due to the unpredictable nature of gameplay and the complex interplay of numerous influencing factors. In this study, we present a deep learning framework that combines a one-dimensional convolutional neural network (1D CNN) with a Transformer architecture to improve prediction accuracy. The 1D CNN effectively captures local spatial patterns in structured match data, while the Transformer leverages self-attention mechanisms to model long-range dependencies. This hybrid design enables the model to uncover nuanced feature interactions critical to outcome prediction. We evaluate our approach on a benchmark sports dataset, where it outperforms traditional machine learning methods and standard deep learning models in both accuracy and robustness. Our results demonstrate the…
Comparison results with machine learning models
As shown in Table 4, our proposed model outperformed all baseline approaches in result prediction of sports events. Specifically, it achieved the best accuracy of 54.73% and MCC of 0.2666, demonstrating its ability to distinctly capture major patterns in the data and enhance generalization. Tree-based models exhibited varying performance levels. The decision tree model, prone to overfitting, recorded the lowest accuracy among tree-based methods at 41.31% with an MCC of 0.0910. In comparison, random forest improved upon this, achieving an accuracy of 50.20% and an MCC of 0.1875. The ensemble nature of random forest mitigates overfitting and enhances model generalization; however, its limitations in capturing strong feature relationships hinder further performance gains. Likewise, the \documentclass[12pt]{minimal} \usepackage{amsmath}…
Comparison results with deep learning models
As described in Table 5, different deep learning architectures demonstrate varying levels of predictive performance in sports event outcome prediction. Among the models evaluated, our proposed approach achieved the highest performance across all metrics, underscoring its capability to effectively capture complex relationships within the dataset. MLP exhibited the lowest ACC of 40.24% and an MCC of 0.0736, indicating its limited ability to generalize well for this task. This suggests that while MLP is effective for tabular data, it may struggle to capture intricate dependencies without extensive hyperparameter tuning. RNNs outperformed MLP, achieving an ACC of 49.39% and an MCC of 0.1745. This aligns with expectations, as RNNs are designed to model sequential dependencies, which can be beneficial in sports event data with temporal…
Editorial reading guide
- What data created and tested the model, and was evaluation independent?
- What is the cost of a false positive, false negative or out-of-distribution decision?
