2016 Fiscal Year Final Research Report
Temporal-Spatial Community Extraction in Social Media
Project/Area Number |
26330345
|
Research Category |
Grant-in-Aid for Scientific Research (C)
|
Allocation Type | Multi-year Fund |
Section | 一般 |
Research Field |
Web informatics, Service informatics
|
Research Institution | Wakayama University |
Principal Investigator |
|
Project Period (FY) |
2014-04-01 – 2017-03-31
|
Keywords | 時空間コミュニティ / コミュニティ抽出 / ネットワーク分析 / トピックモデル / Twitter |
Outline of Final Research Achievements |
We study methods to detect and analyze discussion topics created by user communication on social media.. In order to detect temporal-spatial bursts in user communication, we propose dual-graph SR (Spectral Relaxation) method to extract core potions of a dual graph, which is created from replies on Twitter with specified time interval threshold. Furthermore, we propose a method to classify topics for bag-of-words, which is created with word 2-grams instead of words, by using a generative topic model such as latent Dirichlet allocation (LDA) and visualize the results as topic sequences, which means time series variation of topics, or a topic graph, which means word relationships in each topic.
|
Free Research Field |
Webマイニング
|