RESUMO
Objective To analyze the medication regularity of TCM famous doctors for diabetes by using improved Apriori algorithm to obtain more efficient data mining methods. Methods This article put forward Apriori vertical data storage, and improved ADPM obtained by difference set method was used to conduct data mining to find out the medication regularity in Guo Jia Ji Ming Lao Zhong Yi Tang Niao Bing Yan An Liang Fang. Results After screening, 402 prescriptions were included, involving 24 kinds of high-frequency medicine, 15 high-frequency medical combinations, and 18 highly-dependent medical combinations, which were mainly tonifying deficiency medicine, clearing heat medicine, blood-activating and stasis-resolving medicine, and damp-draining diuretic medicine. Conclusion ADPM algorithm can be applied in the analysis on medication regularity and find out high-frequency medicine, medical combinations and medical dependent relation, with high efficiency.