Abstract:Frequent itemsets mining is often used to find medication law from dataset of Chinese herb prescriptions. Threshold of support count is difficult to set for traditional algorithm of frequent itemsets mining. In the meantime, the number of frequent itemsets is always so big that the result is hard to understand. Some algorithms were proposed to find significant and redundant-aware itemsets. However, the itemsets obtained could not reflect all the information in the dataset. In this paper, a new method was proposed to obtain a collection of itemsets which had the feature of significant, redundant-aware and comprehensive. Firstly, closed frequent itemsets were mined from the dataset of Chinese herbs prescriptions using CHARM algorithm. Then, the itemsets were compressed by FICMC (Frequent Itemsets Compressing based on Minimum Cover) algorithm. Medication law of Chinese herbs could be fully mined from the dataset using this method.

Frequent itemsets compressing based on minimum cover: An efficient method for mining medication law of Chinese herbs