Abstract | ||
---|---|---|
This paper proposes a quality topic extraction on Twitter based on author's role on bipartite networks. We suppose that author's role which means who were in what group, affects the quality of extracted topics. Our proposed method expresses relations between authors and words as bipartite networks, explores author's role by forming clusters using our original community detection technique, and finds quality topics considering the semantic accuracy of words and author's role. |
Year | DOI | Venue |
---|---|---|
2017 | 10.1007/978-3-319-67786-6_17 | DISCOVERY SCIENCE, DS 2017 |
Keywords | Field | DocType |
Topic extraction, Social media analysis, Twitter analysis, Bipartite network, Data mining, Community detection | World Wide Web,Information retrieval,Computer science,Bipartite graph | Conference |
Volume | ISSN | Citations |
10558 | 0302-9743 | 0 |
PageRank | References | Authors |
0.34 | 7 | 4 |
Name | Order | Citations | PageRank |
---|---|---|---|
Takako Hashimoto | 1 | 50 | 18.47 |
Tetsuji Kuboyama | 2 | 140 | 29.36 |
Hiroshi Okamoto | 3 | 1 | 2.75 |
Shin, K. | 4 | 13 | 10.86 |