| dc.description.abstract | In modern medicine, research on traditional Chinese medicine (TCM) is increasing, with numerous studies attempting to verify the efficacy of TCM herbs for specific diseases. If we can infer the correlation between TCM herb combinations and diseases, it could significantly reduce costs in the field of TCM research. This study aims to analyze the association between common herb combinations and disease classifications in the ancient Chinese medical text ＂Pu-ji-Fang(普濟方)＂ using data mining techniques. We will integrate text mining, association analysis, statistical analysis, and case studies based on current herbal research literature.First, we will process the extracted text data from ＂Pu-ji-Fang＂ to filter out eligible prescriptions and herb keywords. Second, we will apply association analysis to identify frequent co-occurrences in the data, revealing potential relationships between symptoms and herb combinations. Third, we will use chi-square tests to evaluate the significance of the discovered associations. Finally, we will compare and verify these findings with current research literature.
Using the ＂Xiaoke(消渴)＂ section in ＂Pu-ji-Fang＂ as an example, ＂Xiaoke＂ is considered to describe symptoms of late-stage diabetes in modern medicine. Through the aforementioned data mining process, we identified several herb combinations significantly associated with ＂Xiaoke.＂ We then selected the more common combinations and reviewed the literature, further confirming that modern medical research supports the effectiveness of TCM prescriptions containing these herb combinations for treating diabetes. This process provides a feasible method for systematically studying TCM classics, expanding the research approach of TCM literature in modern medicine. | en_US |