Association rule mining and network analysis in oriental medicine

Area of Science:

• Traditional East Asian Medicine
• Medical Informatics
• Data Mining

Background:

• Extracting meaningful patterns from large text datasets is increasingly vital.
• Oriental medicine texts contain extensive prescription data linking symptoms to treatments.
• Understanding these relationships can improve the application of traditional remedies.

Purpose of the Study:

• To analyze prescription data from the Bangyakhappyeon, an oriental medicine text.
• To identify and map relationships between categorized symptoms and herbal medicines.
• To leverage data mining techniques for insights into traditional pharmacotherapy.

Main Methods:

• Utilized an association rule algorithm to discover co-occurring patterns.
• Employed network analysis to visualize and understand symptom-medicine connections.
• Processed prescription data covering approximately 54 categorized symptoms.

Main Results:

• Identified significant and informative relationships between specific symptoms and herbal medicines.
• The analysis revealed patterns not easily discernible through traditional review.
• Established a data-driven foundation for understanding traditional herbal treatments.

Conclusions:

• Data mining techniques, specifically association rules and network analysis, are effective for analyzing traditional medicine prescriptions.
• The study successfully mapped symptom-medicine relationships within the Bangyakhappyeon.
• Findings contribute to a deeper, evidence-based understanding of oriental medicine practices.