The importance of graph databases and graph learning for clinical applications

Area of Science:

• Clinical Informatics
• Data Science
• Database Management

Background:

• Clinical data is growing in volume and complexity, challenging traditional relational database storage and analysis.
• Tabular data structures hinder the effective management and retrieval of interlinked clinical information.

Approach:

• This survey reviews state-of-the-art graph database management systems and graph learning algorithms.
• Graph databases store data as nodes and edges, facilitating analysis through graph learning.
• Graph learning encompasses representation learning and analytics for tasks like classification and link prediction.

Key Points:

• Graph databases are adept at handling interconnected clinical data.
• Graph learning provides methods for analyzing complex data structures.
• Applications span visualization, classification, and link prediction in clinical domains.

Conclusions:

• Graph databases and learning offer a robust framework for clinical data challenges.
• This approach enhances the potential for solving domain-specific problems in healthcare.
• A detailed use case is provided for understanding graph learning algorithms.