Exploring significant interactions in live news
News monitoring is of interest to detect current news and track developing stories, but also to explore what is being talked about. In this article, we present an approach to monitoring live feeds of news articles and detecting significant (co-)occurrences of terms compared to a learning background c...
Saved in:
| Main Authors: | , , |
|---|---|
| Format: | Chapter/Article Conference Paper |
| Language: | English |
| Published: |
2018
|
| In: |
NewsIR: recent trends in news information retrieval
|
| Online Access: | Verlag, H, Volltext: http://ceur-ws.org/Vol-2079/paper9.pdf
|
| Author Notes: | Erich Schubert, Andreas Spitz, Michael Gertz |
| Summary: | News monitoring is of interest to detect current news and track developing stories, but also to explore what is being talked about. In this article, we present an approach to monitoring live feeds of news articles and detecting significant (co-)occurrences of terms compared to a learning background corpus. We visualize the result as a graph-structured semantic word cloud that uses a stochastic neighbor embedding (SNE) based layout and visualizes edges between related terms. We give visual examples of our prototype that processes news as they are crawled from dozens of news sites. |
|---|---|
| Item Description: | Gesehen am 19.02.2019 |
| Physical Description: | Online Resource |