MSE for Algorithms Trained on Datasets with and without Additional Features, Second Method_ +F indicates training algorithms with the additional news and injury features_
| Players Position | Ridge | Ridge+F | LightGBM | LightGBM+F | CNN (Frees et al., 2024) | CNN+F | LSTM | LSTM+F |
|---|---|---|---|---|---|---|---|---|
| GK | 7.61 | 7.61 | 7.54 | 7.47 | 7.34 | 4.89 | 3.61 | 3.61 |
| DEF | 5.51 | 5.51 | 5.51 | 5.51 | 5.71 | 5.51 | 7.03 | 6.97 |
| MID | 6.05 | 6.06 | 5.95 | 5.94 | 5.46 | 5.92 | 7.02 | 7.08 |
| FWD | 9.09 | 9.07 | 9.21 | 9.24 | 6.56 | 6.20 | 10.16 | 11.20 |
| Average | 7.06 | 7.06 | 7.05 | 7.04 | 6.27 | 5.63 | 6.95 | 7.21 |
Performance of Algorithms Trained on Datasets with and witout News Sentiments, Classification Task_
| Algorithm/First Method | Accuraccy% | F1 score% |
|---|---|---|
| Baseline | 86.22 | 82.35 |
| Catboost | 84.45 | 76.59 |
| Catboost+News Sentiment | 85.15 | 77.89 |
| XGBoost | 83.74 | 75.78 |
| XGBoost+News Sentiment | 84.09 | 75.93 |
| GB | 84.09 | 76.43 |
| GB+News Sentiment | 83.39 | 74.86 |
Comparison of proposed methods in Related Works_
| Refrence | EPL Data | Textual Data | Other Metadata | ML-based | DL-based |
|---|---|---|---|---|---|
| Bangdiwala et al., 2022; Hermann and Ntoso, 2015; Rajesh et al., 2022; Shah et al., 2023 | ✓ | ✗ | ✗ | ✓ | ✗ |
| Gupta, 2019; Lombu et al., 2024; Ramdas, 2022 | ✓ | ✗ | ✗ | ✓ | ✓ |
| Bonello et al., 2019 | ✓ | ✓ | ✓ | ✓ | ✗ |
| Baughman et al., 2021; Frees et al., 2024, Our Model | ✓ | ✓ | ✓ | ✓ | ✓ |
Performance of Algorithms Trained on Datasets with and without News Sentiments, Regression Task_
| Algorithm/First Method | MSE | RMSE |
|---|---|---|
| Catboost | 7.27 | 2.69 |
| Catboost+News Sentiment | 7.51 | 2.74 |
| XGBoost | 7.46 | 2.73 |
| XGBoost+News Sentiment | 7.40 | 2.72 |
| GB | 7.37 | 2.71 |
| GB+News Sentiment | 7.74 | 2.78 |
Players’ Presence Prediction Using Catboost for Different Postions, with and witout News Sentiment_
| Players Position | Accuraccy% | F1 score% |
|---|---|---|
| GK | 87.87 | 74.99 |
| GK+News Sentiment | 95.23 | 87.71 |
| DEF | 77.52 | 73.68 |
| DEF+News Sentiment | 80.00 | 73.84 |
| MID | 86.66 | 80.85 |
| MID+News Sentimen | 92.66 | 88.57 |
| FWD | 76.92 | 62.50 |
| FWD+News Sentiment | 92.50 | 82.35 |