Table 3.
The results of 5-fold cross-validations with the 1–3-gram language model.
| CV 1 | CV 2 | CV 3 | CV 4 | CV 5 | Average | |
|---|---|---|---|---|---|---|
| LR | 0.9545 | 0.9535 | 0.9525 | 0.9540 | 0.9540 | 0.9537 |
| RF | 0.9403 | 0.9457 | 0.9452 | 0.9437 | 0.9374 | 0.9424 |
| Multinomial NB | 0.9183 | 0.9202 | 0.9207 | 0.9154 | 0.9168 | 0.9183 |
| MLP | 0.9310 | 0.9300 | 0.9344 | 0.9315 | 0.9281 | 0.9310 |
| KNN | 0.8127 | 0.8112 | 0.8136 | 0.8078 | 0.8083 | 0.8107 |
| SVM | 0.9310 | 0.9315 | 0.9339 | 0.9334 | 0.9325 | 0.9325 |
| XGBoost | 0.9618 | 0.9569 | 0.9574 | 0.9584 | 0.9603 | 0.9590 |
CV, cross-validation; LR, logistic regression; RF, random forest; NB, naive Bayes; MLP, multi-layer perceptron; KNN, k-nearest neighbor; SVM, support vector machine; XGBoost, extreme gradient boosting.