eCite Digital Repository

UDeKAM: A methodology for acquiring declarative structured knowledge from unstructured knowledge resources


Ali, M and Lee, S and Kang, BH, UDeKAM: A methodology for acquiring declarative structured knowledge from unstructured knowledge resources, Proceedings of the International Conference on Machine Learning and Cybernetics 2016, 10-13 July 2016, Maison Glad Jeju, Jejuleju Island, South Korea, pp. 177-182. ISBN 9781509003891 (2016) [Refereed Conference Paper]

Copyright Statement

Copyright 2016 IEEE.

DOI: doi:10.1109/ICMLC.2016.7860897


An effective knowledge representation has always proved its importance for mankind intelligence. Among various kinds of knowledge, declarative knowledge has a vital role in medical domain and is critical for health-care safety and quality. A large volume of declarative knowledge is hidden in multiple knowledge resources such as clinical notes, standard guidelines etc. that can play an important role in decision support systems as well as in health and wellness applications after structured transformation. In this paper, an Unstructured Declarative Knowledge Acquisition Methodology, called UDeKAM, is proposed that acquires and constructs the declarative structured knowledge from unstructured knowledge resources using Documents Clustering, Topic Modeling, and Controlled Natural Language processing techniques. The proposed methodology is designed for different domains to serve a variety of applications. It is an ongoing work and for the realization of UDeKAM, a diabetes scenario is explained through example.

Item Details

Item Type:Refereed Conference Paper
Keywords:controlled natural language, declarative knowledge, ontological model, resource classification, topic modeling
Research Division:Information and Computing Sciences
Research Group:Artificial intelligence
Research Field:Modelling and simulation
Objective Division:Expanding Knowledge
Objective Group:Expanding knowledge
Objective Field:Expanding knowledge in the environmental sciences
UTAS Author:Kang, BH (Professor Byeong Kang)
ID Code:118192
Year Published:2016
Deposited By:Information and Communication Technology
Deposited On:2017-07-06
Last Modified:2018-01-15

Repository Staff Only: item control page