Ontology modeling of semantics in social media: Public issue knowledge base (PIKB) of the Weibo
ZHOU Yan; LI Wei; YUAN Xingfu; ZHANG Pengyi; Zhang Pengyi (E-mail: pengyi@pku.edu.cn)
2014-03-25
Source PublicationChinese Journal of Library and Information Science
ISSN1674-3393
Volume7Issue:1Pages:16-30
Abstract
Purpose: This study aims to construct an ontology to model the semantics of social media streams, in particular, trending topics and public issues.

Design/methodology/approach: Our knowledge base included 10 public events and topics from Weibo respectively, which were collected through keyword search and a crawler program. We used a semi-automatic approach to model and annotate the semantics in social media, and adapted the multi-layered ontology to refine the design based on previous researches, then we used named entity recognition (NER) to extract entities to instantiate the ontology. Relationships were extracted based on co-occurrence measures. Finally, we manually conducted post-filtering evaluation and edited the extracted entities and relationships.

Findings: An initial assessment demonstrated that our multi-layered ontology supports various types of queries and analyses in the public issue knowledge base (PIKB), which can serve as an effective tool to query, understand and trace public issues.

Research limitations: Manual involvement cannot meet the requirements for challenges of sustainable developments. Since the relationships extracted are fully based on the co-occurrence of entities, rich semantic relationships, such as how much the key players have been involved, could not be fully reflected. Besides, the user evaluation is necessary for further ontology assessment.

Practical implications: The PIKB can be used by regular Web users and policy makers to query, understand, and make sense of public events and topics. The methodology and reusable ontology model are useful for institutions that are interested in making use of the social media data.

Originality/value: In this study, a multi-layered ontology is applied to model the evolving semantics of public events and trending topics in social media, and the semi-automatic approach could make it possible to extract entities and relationships from large amount of unstructured short texts of user generated content (UGC) from social media.
;
Purpose: This study aims to construct an ontology to model the semantics of social media streams, in particular, trending topics and public issues.

Design/methodology/approach: Our knowledge base included 10 public events and topics from Weibo respectively, which were collected through keyword search and a crawler program. We used a semi-automatic approach to model and annotate the semantics in social media, and adapted the multi-layered ontology to refine the design based on previous researches, then we used named entity recognition (NER) to extract entities to instantiate the ontology. Relationships were extracted based on co-occurrence measures. Finally, we manually conducted post-filtering evaluation and edited the extracted entities and relationships.

Findings: An initial assessment demonstrated that our multi-layered ontology supports various types of queries and analyses in the public issue knowledge base (PIKB), which can serve as an effective tool to query, understand and trace public issues.

Research limitations: Manual involvement cannot meet the requirements for challenges of sustainable developments. Since the relationships extracted are fully based on the co-occurrence of entities, rich semantic relationships, such as how much the key players have been involved, could not be fully reflected. Besides, the user evaluation is necessary for further ontology assessment.

Practical implications: The PIKB can be used by regular Web users and policy makers to query, understand, and make sense of public events and topics. The methodology and reusable ontology model are useful for institutions that are interested in making use of the social media data.

Originality/value: In this study, a multi-layered ontology is applied to model the evolving semantics of public events and trending topics in social media, and the semi-automatic approach could make it possible to extract entities and relationships from large amount of unstructured short texts of user generated content (UGC) from social media.
KeywordOntology Knowledge Organization Public Issue Knowledge Base (Pikb) Public Issues Social Media
Subject Area编辑出版
URL查看原文
Funding OrganizationThis work is supported by Beijing Thinker Workshop (Grant No. XK201211001).
Document Type期刊论文
Identifierhttp://ir.las.ac.cn/handle/12502/6818
CollectionJournal of Data and Information Science_Chinese Journal of Library and Information Science-2014
Corresponding AuthorZhang Pengyi (E-mail: pengyi@pku.edu.cn)
Recommended Citation
GB/T 7714
ZHOU Yan,LI Wei,YUAN Xingfu,et al. Ontology modeling of semantics in social media: Public issue knowledge base (PIKB) of the Weibo[J]. Chinese Journal of Library and Information Science,2014,7(1):16-30.
APA ZHOU Yan,LI Wei,YUAN Xingfu,ZHANG Pengyi,&Zhang Pengyi .(2014).Ontology modeling of semantics in social media: Public issue knowledge base (PIKB) of the Weibo.Chinese Journal of Library and Information Science,7(1),16-30.
MLA ZHOU Yan,et al."Ontology modeling of semantics in social media: Public issue knowledge base (PIKB) of the Weibo".Chinese Journal of Library and Information Science 7.1(2014):16-30.
Files in This Item: Download All
File Name/Size DocType Version Access License
2.Yan ZHOU.pdf(3377KB) 开放获取View Download
Related Services
Recommend this item
Bookmark
Usage statistics
Export to Endnote
Google Scholar
Similar articles in Google Scholar
[ZHOU Yan]'s Articles
[LI Wei]'s Articles
[YUAN Xingfu]'s Articles
Baidu academic
Similar articles in Baidu academic
[ZHOU Yan]'s Articles
[LI Wei]'s Articles
[YUAN Xingfu]'s Articles
Bing Scholar
Similar articles in Bing Scholar
[ZHOU Yan]'s Articles
[LI Wei]'s Articles
[YUAN Xingfu]'s Articles
Terms of Use
No data!
Social Bookmark/Share
File name: 2.Yan ZHOU.pdf
Format: Adobe PDF
All comments (0)
No comment.
 

Items in the repository are protected by copyright, with all rights reserved, unless otherwise indicated.