EpiK: A Knowledge Base for Epidemiological Modeling and Analytics of Infectious Diseases.

Area of Science:

• Computational epidemiology
• Data science
• Epidemic modeling

Background:

• Advances in computing and data science enable innovative modeling for epidemiology.
• Large, heterogeneous datasets in computational epidemiology pose significant data management challenges.
• Scalable systems are needed to store, manage, and integrate these complex datasets.

Purpose of the Study:

• To develop EpiK, a knowledge base for epidemic science decision support and analytical environments.
• To create a framework linking input/output datasets for spatio-temporal and social reasoning in epidemic analysis.
• To facilitate planning and intervention strategies before and during epidemics.

Main Methods:

• Developed EpiK, a knowledge base linking modeling workflow data and metadata using a controlled vocabulary.
• Utilized semantic web technologies and RDF to represent datasets, handling schema and location heterogeneity.
• Created a query bank using SPARQL to support diverse questions across the computational epidemiology modeling pipeline.

Main Results:

• EpiK successfully links modeling data and metadata, supporting agent-based and aggregate models.
• The system hides data heterogeneity, enabling efficient queries from model construction to simulation output.
• Query performance is influenced by the underlying hardware and RDF engine.

Conclusions:

• EpiK provides a scalable data management framework for computational epidemiology.
• The knowledge base enhances decision support and analytical capabilities for epidemic science.
• Further research can optimize performance by considering hardware and RDF engine choices.