Have a personal or library account? Click to login
A Deep Learning Approach Based on Interpretable Feature Importance for Predicting Sports Results Cover

A Deep Learning Approach Based on Interpretable Feature Importance for Predicting Sports Results

Open Access
|Mar 2025

Figures & Tables

Figure 1.

The proposed methodology.
The proposed methodology.

Figure 2.

Receiver Operating Characteristic (ROC) curves of the six models.
Receiver Operating Characteristic (ROC) curves of the six models.

Figure 3.

Confusion matrix of the MLP model with values −1 (Away Win), 0 (Draw), and 1 (Home Win).
Confusion matrix of the MLP model with values −1 (Away Win), 0 (Draw), and 1 (Home Win).

Figure 4.

SHAP summary plot.
SHAP summary plot.

Figure 5.

Force plot for the first prediction's explanation.
Force plot for the first prediction's explanation.

Figure 6.

Correlation matrix between features.
Correlation matrix between features.

Description of features_

Feature NameDescription
HomeTeamHome Team
AwayTeamAway Team
FTRFull Time Result (H=Home Win, D=Draw, A=Away Win)
HSHome Team Shots
ASAway Team Shots
HSTHome Team Shots on Target
ASTAway Team Shots on Target
HCHome Team Corners
ACAway Team Corners
HFHome Team Fouls Committed
AFAway Team Fouls Committed
HYHome Team Yellow Cards
AYAway Team Yellow Cards
HRHome Team Red Cards
ARAway Team Red Cards

Performance comparison of the six models in our study_

ClassifiersAccuracyPrecisionRecallF1-scoreAUC
Logistic Regression0.640.620.640.630.81
Naïve Bayes0.600.600.600.600.75
Support Vector Machine0.640.630.640.640.82
Multilayer Perceptron0.670.660.670.670.83
Random Forest0.640.620.640.620.81
XGBoost0.660.630.660.640.82
Language: English
Page range: 56 - 72
Published on: Mar 19, 2025
Published by: International Association of Computer Science in Sport
In partnership with: Paradigm Publishing Services
Publication frequency: 2 issues per year

© 2025 Messaoud Bendiaf, Hakima Khelifi, Djamila Mohdeb, Mouhoub Belazzoug, Abdelhamid Saifi, published by International Association of Computer Science in Sport
This work is licensed under the Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 License.