TCM Prescription Compatibility Based on Improved Association Rules Algorithm

Objective: To analyze the traditional Chinese medicine (TCM) prescription regularity of Treatise on Febrile Diseases by using improved Apriori algorithm to obtain more efficient data mining. Methods: 113 formulae from Treatise on Febrile Diseases were collected and terms of herbs in this prescription were standardized. This paper put forward valid value index storage, and fast intersection operation to improve the efficiency of mining TCM data. The support-confidence-lift framework is adopted to evaluate the effectiveness of the rules and avoid the generation of meaningless rules. Results: 18 high-frequency herbs with occurrence of 10 times or above, including Licorice, Cassia Twig, Jujube and Ginseng, etc. Among18 high-frequency herbs, 52 combinations are obtained classical traditional herb pairs, such as Licorice-Cassia Twig, Jujube-Ginger and Ginger-Licorice, etc. Conclusion: The improved Apriori algorithm can be applied in the analysis on prescription compatibility and find out high-frequency herbs and herbs combinations with low storage consumption and high efficiency. The experimental result can provide references for clinical use of herbs which reveal the compatibility rule of the classical prescription.