Purpose: This study aims to analyze the co-occurrence ofpathological symptoms and corresponding acupoints as documentedby the comprehensive acupuncture and moxibustionrecords in the classical texts of Far East traditional medicine asan aid to a more efficient understanding of the tacit treatmentprinciples of ancient physicians.Methods: The Classic of Supporting Life with Acupunctureand Moxibustion (Zhenjiu Zisheng Jing; hereinafter ZZJ),which contains the largest amount of treatment cases andleft a strong impact on the Far East medical history, wasselected as the primary reference book for the analysis.ZZJ was first digitized and co-occurring pathology-acupointpairs were extracted and preprocessed into an analyzableformat. The pathology-acupoint co-occurrence analysis wasperformed by applying 5 values of set-theoretic measures(weighted Euclidean distance, Canberra distance, Euclideandistance, chi-squared distance, and Jaccard similarity), whichmeasure the distance between the observed and expectedco-occurrence counts, and 2 values of probabilistic measures(association strength and Fisher’s exact test), which measurethe probability of observed co-occurrences. The analysisresults were used for a prediction simulation in order to measureand compare the extent to which pathologies can bepredicted from acupoints.Results: The treatment records contained in ZZJ werepreprocessed, which yielded 4162 pathology-acupoint sets.Co-occurrence was performed applying 7 different analysisvariables, followed by a prediction simulation. The predictionsimulation results revealed the weighted Euclidean distance had the highest prediction rate with 24.31%, followed by Canberradistance (23.14%) and association strength (21.29%).Conclusion: The weighted Euclidean distance among theset-theoretic measures and the association strength amongthe probabilistic measures were verified to be the most efficientanalysis methods in analyzing the correlation betweenacupoints and pathologies found in the classical medical texts.
展开▼
