Web Ontology Engineering and Applications: A Case Study in Traditional Chinese Medicine

1
Web Ontology Engineering and Applications A
Case Study in Traditional Chinese Medicine
Yuxin Mao CCNT Lab of Zhejiang University
2
Outline

• Background
• Ontology Modeling
• Ontology Development
• Ontology Reuse
• Applications

3
Background

• Ontologies are the specification of
  conceptualizations, used for programs and humans
  to share knowledge.
• Semantic Web facilitates the engineering of
  various large-scale online ontologies
• UMLS for integrating biomedical terminology
• Gene Ontology for gene product and
• MGED Ontology for microarray experiment

4
Background(2)

• TCM Research requires a great deal of knowledge
  from different sources
• TCM herbal medicine, TCM diseases, TCM EMRs, TCM
  Literatures
• The size of existing TCM knowledge base is too
  large for any TCM scientist or doctor to handle.
• A large-scale ontology is critical in the sharing
  and integration of information and knowledge in
  TCM
• to overcome the problem of semantic heterogeneity
  
• and encode domain knowledge in reusable format

5
Ontology Modeling

• Here we present the upper-level framework for
  modeling the TCM ontology.
• The current TCM ontology contains 12 major
  categories for each sub-domain.
• The TCM ontology includes 2 components
• concept system
• semantic system

6
Semantic System

• We have defined 59 semantic relationships
  between content classes and 104 TCM semantic
  types
• According to the time, function, space, entity
  and class attributes of TCM domain knowledge
• plus all the semantic types from UMLS

7
Concept System

• The concept system defines the concepts in TCM
  domain, and it contains content classes and
  implemental classes.
• Content class represents the concrete domain
  knowledge of the TCM discipline.
• In order to unify class structure, we formulize
  key properties and relationships of content class
  as implemental classes
• Name Class represents various name terms (e.g.
  alias, English name, or synonym) of an instance
• Definition Class represents the scientific
  definition of a in-stance
• Explanation Class represents additional
  explanations to instance definition
• Relation Class represents the relationship
  between two instances

8
Ontology Development

• Requirements of the Ontology development
  environment (ODE) for TCM Ontology
• Facilitate professional ontology development
• The TCM ontology is developed collaboratively by
  experts from several institutions, so the ODE
  should enable remote access and modification on
  the ontology
• The TCM ontology is divided into several
  categories both logically and geographically, so
  the ODE should enable ontology integration and
  federation in a distributed environment like the
  Web
• We have built a Web-based ODE, called TOtE (TCM
  Ontology Editor)

9
Web-based Ontology Editor

• TOtE is an editor that allows users to edit and
  explore ontology online
• TOtE runs on the server-side and publishes
  large-scale TCM ontologies to users through Web
  services
• Users can browse and edit TCM ontologies anywhere
  with their Web browsers.
• Incorporates a back-end database for ontology
  storage
• Compatible with popular ontology formats

10
Web-based Ontology Editor(2)

• The user interface incorporates an open-source
  AJAX framework
• a tree-based view for classes and a form-based
  view for instances of an ontology, similar with
  Protégé
• employ a layered privilege mechanism in TOtE and
  users that play different roles in the process of
  ontology development hold different privileges

11
Development Procedure

• Analyze and determine knowledge sources
• Construct upper-level conceptual framework
• Determine and assign developing tasks
• Extend conceptual hierarchy
• Materialize ontology contents
• Check and revise contents
• Publish ontology user interface

12
Major Result

• there have been more than 20,000 classes and
  100,000 instances defined in the current
  knowledge base of the TCM ontology
• The ontology under development is still part of
  the complete one
• The TCM ontology has become a distributed
  large-scale knowledge base for TCM domain
  knowledge and can be accessed on-line

13
Ontology Reuse

• Locality of knowledge reference
• A typical application can only use a small
  proportion of the TCM Ontology
• Sub-ontologies are context-specific portions of
  the TCM ontology
• Sub-ontology cache can hold SubOs as cache blocks
  for ontology reuse.

14
Applications

• Database Integration
• DartGrid a semantic-based database grid platform
  
• we adopt the semantics of the TCM ontology to
  integrate massive amount of TCM databases based
  on the Grid infrastructure
• Informatics Search Engine
• DartSearch a semantic-based search engine for
  TCM informatics
• Provides users with a Google-like search
  interface to perform semantic search
• We harness the ontology semantics to improve
  searching experience in DartSearch.

15
Conclusion

• Ontology engineering is a key issue for realizing
  the Semantic Web vision.
• we propose a comprehensive approach to model,
  develop, and reuse large-scale Web ontology for
  the TCM discipline.
• The scale of the TCM ontology has reached 20,000
  classes and 100,000 instances by now. The TCM
  ontology has been used to support several
  semantic-based applications for TCM.

16
Thank You
QA