Abstract | ||
---|---|---|
•We motivate event discovery from geo-located tweets in Twitter.•We propose to tackle this problem through density-based clustering with noise.•We formulate Tweet-SCAN within GDBSCAN to cope with Twitter objects.•We demonstrate how Tweet-SCAN is able to discover real-world events.•We show the benefits of considering the textual component of a geo-located tweet. |
Year | DOI | Venue |
---|---|---|
2017 | 10.1016/j.patrec.2016.08.010 | Pattern Recognition Letters |
Keywords | DocType | Volume |
Twitter,Unsupervised learning,Event discovery,DBSCAN,Probabilistic topic models,Hierarchical Dirichlet Process (HDP) | Journal | 93 |
Issue | ISSN | Citations |
93 | 0167-8655 | 5 |
PageRank | References | Authors |
0.43 | 13 | 4 |
Name | Order | Citations | PageRank |
---|---|---|---|
Joan Capdevila | 1 | 19 | 1.68 |
Jesús Cerquides | 2 | 200 | 16.84 |
Jordi Nin | 3 | 311 | 26.53 |
Jordi Torres | 4 | 1330 | 80.21 |