Using different linear models I tried to predict the winner of a tennis matches based on 2 features.
Sample data:
Since the only feautres available for predicting are the personal scores of the players (lk1 & lk2),
plotting them reveals a linear relationship
sns.lineplot(x = 'lk1', y = 'lk2', data = dfs, hue = 'match_outcome')
plt.show()
I build and trained the following linear model classifiers:
- Ridge
- Lasso
- KNN
- Logistic Regression
- Decision Tree
Then, tested them on the new
and the results I've got:
| date_of_match | player1_id | player1_name | player2_id | player2_name | lk1 | lk2 | final_outcome | Ridge - L2 | SVC | KNN | LogReg | DecisionTree | |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 0 | 27.09.2020 | 20754594 | Fischer | 29753705 | Pfau | 15.1 | 15.1 | lose | win | lose | lose | win | lose |
| 1 | 27.09.2020 | 20754594 | Fischer | 20355195 | Kösters | 15.1 | 20.1 | win | win | win | win | win | win |
| 2 | 13.09.2020 | 20754594 | Fischer | 28803686 | Opalka | 15.1 | 22.1 | win | win | win | win | win | win |
| 3 | 05.09.2020 | 20754594 | Fischer | 20457747 | Goedecke | 15.1 | 15.1 | lose | win | lose | lose | win | lose |
| 4 | 23.08.2020 | 20754594 | Fischer | 20357314 | van Raay | 15.1 | 23.1 | win | win | win | win | win | win |
| 5 | 16.08.2020 | 20754594 | Fischer | 20263300 | Verheyen | 15.1 | 23.1 | win | win | win | win | win | win |
| 6 | 09.08.2020 | 20754594 | Fischer | 29050113 | Siebert | 15.1 | 22.1 | win | win | win | win | win | win |
| 7 | 12.07.2020 | 20754594 | Fischer | 20651971 | Menze | 15.1 | 11.7 | lose | lose | lose | lose | lose | win |
| 8 | 09.07.2020 | 20754594 | Fischer | 20551904 | Bodem | 15.1 | 15.1 | lose | win | lose | lose | win | lose |
| 9 | 09.07.2020 | 20754594 | Fischer | 20653485 | Rajapreyar | 15.1 | 8.0 | lose | lose | lose | lose | lose | lose |
| 10 | 29.06.2020 | 20754594 | Fischer | 20556878 | Grigorieva | 15.1 | 9.7 | lose | lose | lose | lose | lose | lose |
| 11 | 25.06.2020 | 20754594 | Fischer | 20652120 | Götz | 15.1 | 19.1 | win | win | win | win | win | win |
| 12 | 25.06.2020 | 20754594 | Fischer | 20651858 | Kriegbaum | 15.1 | 21.8 | win | win | win | win | win | win |
| 13 | 25.06.2020 | 20754594 | Fischer | 20653608 | Hein | 15.1 | 13.1 | lose | lose | lose | lose | win | lose |
| 14 | 21.06.2020 | 20754594 | Fischer | 20652121 | Pohl | 15.1 | 16.8 | win | win | win | win | win | win |
| 15 | 19.06.2020 | 20754594 | Fischer | 20651901 | Schaefer | 15.1 | 6.1 | lose | lose | lose | lose | lose | lose |
| 16 | 12.06.2020 | 20754594 | Fischer | 20652120 | Götz | 15.1 | 19.1 | win | win | win | win | win | win |
| 17 | 11.06.2020 | 20754594 | Fischer | 20352348 | Kroll | 15.1 | 8.1 | lose | lose | lose | lose | lose | lose |
| 18 | 24.02.2020 | 20754594 | Fischer | 20651370 | Wolter | 15.1 | 12.1 | lose | lose | lose | win | lose | win |
| 19 | 24.02.2020 | 20754594 | Fischer | 20655872 | Bongardt | 15.1 | 19.0 | win | win | win | win | win | win |
| 20 | 11.02.2020 | 20754594 | Fischer | 20652388 | Buß | 15.1 | 4.1 | lose | lose | lose | lose | lose | lose |
| 21 | 11.02.2020 | 20754594 | Fischer | 20650020 | Kuhlwein von Rathenow | 15.1 | 13.6 | lose | lose | lose | lose | win | lose |