The goal of this task is to classify tweets into three levels of polarity: positive, negative and neutral. Tweets are written in different variants of the Spanish language: (ES-Spain, CR-Costa Rica, PE-Peru, UY-Uruguay and MX-Mexico).
Publication
Manuel Carlos Díaz-Galiano, Manuel García-Vega, Edgar Casasola, Luis Chiruzzo, Miguel García-Cumbreras, Eugenio Martínez Cámara, Daniela Moctezuma, Arturo Montejo Ráez, Marco Antonio Sobrevilla Cabezudo, Eric Tellez, Mario Graff, Sabino Miranda (2019) Overview of TASS 2019: One More Further for the Global Spanish Sentiment Analysis Corpus. Proceedings of the Iberian Languages Evaluation Forum (IberLEF 2019)
Competition
Language
Spanish (Costa Rica)
Spanish (Mexico)
Spanish (Peru)
Spanish (Spain)
Spanish (Uruguay)
NLP topic
Abstract task
Year
2019
Publication link
Ranking metric
Macro F1
Task results
| System | MacroPrecision | MacroRecall | MacroF1 Sort ascending |
|---|---|---|---|
| ELiRF-UPV | 0.4970 | 0.5360 | 0.5150 |
| RETUYT-InCo | 0.5880 | 0.4540 | 0.5120 |
| ELiRF-UPV | 0.4900 | 0.5120 | 0.5010 |
| Atalaya | 0.4980 | 0.4990 | 0.4990 |
| ELiRF-UPV | 0.4980 | 0.4930 | 0.4960 |
| GTH-ETSIT-UPM | 0.5210 | 0.4660 | 0.4920 |
| GTH-ETSIT-UPM | 0.4970 | 0.4770 | 0.4870 |
| RETUYT-InCo | 0.4870 | 0.4850 | 0.4860 |
| Atalaya | 0.5330 | 0.4440 | 0.4840 |
| Atalaya | 0.4720 | 0.4670 | 0.4690 |
