| Revista: | Computación y sistemas |
| Base de datos: | |
| Número de sistema: | 000560333 |
| ISSN: | 1405-5546 |
| Autores: | Rudrapal, Dwijen1 Das, Amitava2 Bhattacharya, Baby1 |
| Instituciones: | 1National Institute of Technology, Agartala. India 2Indian Institute Of Information Technology, Sricity, Uttar Pradesh. India |
| Año: | 2019 |
| Periodo: | Abr-Jun |
| Volumen: | 23 |
| Número: | 2 |
| Paginación: | 491-500 |
| País: | México |
| Idioma: | Inglés |
| Tipo de documento: | Artículo |
| Resumen en inglés | A Twitter event contains enormous number of tweets including views or comments, real-time update of the event, conversation and often lots of irrelevant information too. Thus, ranking of the most relevant and important tweets for an event is a difficult research problem. In last decade, researchers proposed many state-of-the-art tweet ranking solutions. Performance of these solutions are event based and often return pointless tweets also which deluge the informative tweets. This paper proposes an approach to rank the most relevant and informative tweets for an event. We introduce new features in addition of state-of-the-art features to measure relevance and informativeness of tweets more accurately. The performance of the proposed ranking approach is evaluated through experimental result and reports comparable performance in this domain. |
| Disciplinas: | Ciencias de la computación |
| Palabras clave: | Inteligencia artificial |
| Keyword: | Social media text, Twitter event, Learning to rank, Summarization, Tweet ranking, Artificial intelligence |
| Texto completo: | Texto completo (Ver HTML) Texto completo (Ver PDF) |
