download0 view787
twitter facebook

공공누리This item is licensed Korea Open Government License

Title
Building a Business Knowledge Base by a Supervised Learning and Rule-Based Method
Author(s)
신성호이문용정한민
Publication Year
2015-01-31
Abstract
Natural Language Question Answering (NLQA) and Prescriptive Analytics (PA) have been identified as innovative, emerging technologies in 2015 by the Gartner group. These technologies require knowledge bases that consist of data that has been extracted from unstructured texts. Every business requires a knowledge base for business analytics as it can enhance companies’ competitiveness in their industry. Most intelligent or analytic services depend a lot upon on knowledge bases. However, building a qualified knowledge base is very time consuming and requires a considerable amount of effort, especially if it is to be manually created. Another problem that occurs when creating a knowledge base is that it will be outdated by the time it is completed and will require constant updating even when it is ready in use. For these reason, it is more advisable to create a computerized knowledge base. This research focuses on building a computerized knowledge base for business using a supervised learning and rule-based method. The method proposed in this paper is based on information extraction, but it has been specialized and modified to extract information related only to a business. The business knowledge base created by our system can also be used for advanced functions such as presenting the hierarchy of technologies and products, and the relations between technologies and products. Using our method, these relations can be expanded and customized according to business requirements.
Keyword
Information extraction; business knowledge base; structural support vector machine; named entity recognition; relation extraction
Journal Title
KSII Transactions on Internet and Information Systems
Citation Volume
9
ISSN
1976-7277
Files in This Item:
There are no files associated with this item.
Appears in Collections:
7. KISTI 연구성과 > 학술지 발표논문
URI
https://repository.kisti.re.kr/handle/10580/14447
http://www.ndsl.kr/ndsl/search/detail/article/articleSearchResultDetail.do?cn=JAKO201508449474247
Fulltext
 http://www.ndsl.kr/ndsl/commons/util/ndslOriginalView.do?dbt=JAKO&cn=JAKO201508449474247
Export
RIS (EndNote)
XLS (Excel)
XML

Browse