The task aims at encouraging research on Online Reputation Management. Using as input a set of Twitter profiles, the tasks consists on finding out which authors have more reputational influence (who the influencers or opinion makers are) and which profiles are less influential or have no influence at all. For a given domain (e.g., automotive or banking), the systems’ output is a ranking of profiles according to their probability of being an opinion maker with respect to the concrete domain, optionally including the corresponding weights.
Publication
Enrique Amigó, Jorge Carrillo de Albornoz, Irina Chugur, Adolfo Corujo, Julio Gonzalo, Edgar Meij, Damiano Spina (2014) Overview of RepLab 2014: Author Profiling and Reputation Dimensions for Online Reputation Management.CEUR Proceedings. http://ceur-ws.org/Vol-1180/#CLEF2014wn-Rep-AmigoEt2014
Language
Spanish
NLP topic
Abstract task
Dataset
Year
2014
Publication link
Task results
| System | Accuracy Sort ascending |
|---|---|
| UTDBRG_AR_4 | 0.5700 |
| LyS_AR_1.txt | 0.5600 |
| UTDBRG_AR_1 | 0.5500 |
| UTDBRG_AR_5 | 0.5000 |
| UTDBRG_AR_3 | 0.5000 |
