Table 4.
Classification Accuracy of the Models using LIAR dataset.
| Author and Year | Features | Representation | Classifier | F1 Score |
|---|---|---|---|---|
| Wang [31], 2017 | Text | Word2Vec | Hybrid CNN | 27.01% |
| Long [15], 2017 | Speaker profile | Weighted vectors | LSTM attention | 41.5% |
| Goldani, Safabakhsh [32], 2021 | Text | Glove.6B.300d | CNN with margin loss | 40.58% |
| Goldani, Momtazi [23], 2021 | Text | Word2Vec | CNN | 39.50% |
| Hakak, Alazab [30], 2021 | derived (Text + NER) |
Statistical | RF | 44.15% |
| Samadi, Mousavian [22], 2021 | 2021 | Funnel | CNN | 48.64% |
| The proposed model | Text | TF–IDF-IG | SDL–MLP | 51.05% |