Table 3 Data splitting for PolitiFact and GossipCop datasets (train-validation-test).
From: Towards improved fake news detection using a hybrid RoBERTa and metadata enhanced XGBoost model
Dataset | Train (70%) | Test (20%) | Validation (10%) |
|---|---|---|---|
PolitiFact | |||
 Fake news | 331 | 95 | 47 |
 Real news | 558 | 159 | 80 |
 Total | 889 | 254 | 127 |
GossipCop | |||
 Fake news | 3,734 | 1,067 | 534 |
 Real news | 11,773 | 3,364 | 1,681 |
 Total | 15,507 | 4,431 | 2,215 |
