Objective: It laid the foundation of the large sample data mining for a comprehensive summary concerning four properties theory of traditional Chinese medicine (TCM), and also provided theory clues on determination of four properties for the new resource development of TCM and the clinical use of Chinese medicine.
Method: Four properties data of 8 980 Chinese medicines from "Chinese herbal medicine (CHM)" and associated function index data were chose as data sets. Then, the IBM SPSS Clementine 14.1 data mining platform and Apriori model were adopted to mining classification-association rules, setting the minimum support threshold of rule antecedent and the minimum confidence threshold as 0.5% and 60%.
Result: 11 classification-association rules involved in warm, cold and mild natures were found.
Conclusion: It was discovered that the TCM with functions of "dispelling wind-cold, warming the middle, stopping pain and expelling wind-removing dampness, tonifying kidney yang, unblocking meridians and expelling wind-removing dampness, dispelling cold to stop pain, strengthening sinews-bones and expelling wind-removing dampness" was likely warm-natured, with functions of "tonifying the lung" was likely mild-natured, and with functions of " clearing heat and drying dampness, clearing heat and purging fire, eliminating restlessness" was likely cold-natured.