From: Detecting political biases of named entities and hashtags on Twitter
Results (accuracy; \(F_{1}\)) | Polarized PEM | Complete PEM |
---|---|---|
Using \(\mathbf{z}^{(p)}\) of All Tokens | 0.8369; 0.8366 | 0.8337; 0.8334 |
Using \(\mathbf{z}^{(p)}\) of Only Entities and Hashtags | 0.8339; 0.8337 | 0.8338; 0.8330 |