Knowledge Discovery from Biomedical Ontologies in Cross Domains.

Area of Science:

• Biomedical Informatics
• Data Science

Background:

• Growing demand for medical data sharing and integration in biomedical research.
• Need for semantic interoperability to improve healthcare systems.
• Challenges in achieving semantic interoperability due to underutilization of conceptual models.

Purpose of the Study:

• Propose a semantic framework, Medical Knowledge Discovery and Data Mining (MedKDD), for topic hierarchy construction and semantic interoperability.
• Facilitate knowledge discovery across multiple biological and medical ontologies.

Main Methods:

• Focus on discovering semantic patterns in heterogeneous information networks representing biological ontologies.
• Create a topic hierarchy by analyzing discovered patterns.
• Cluster information networks into topic graphs for cross-domain knowledge discovery.
• Utilize 9 primary biological ontologies from Bio2RDF for a case study.

Main Results:

• Demonstrated cross-domain knowledge discovery using the MedKDD framework.
• Compared MedKDD's effectiveness against the SLAP cross-domain query processing approach.
• Confirmed MedKDD's effectiveness in knowledge discovery from multiple medical ontologies.

Conclusions:

• The MedKDD framework effectively addresses semantic interoperability challenges in biomedical research.
• Discovered semantic patterns significantly contribute to cross-ontology knowledge discovery.
• MedKDD provides a robust approach for integrating and analyzing diverse medical data.