Study on the Compatibility Law of Lobelia Based on Association Rule and Systematic Cluster Analysis

Abstract

Objective: based on association rules and cluster analysis explore the compatibility law of lobelia. Method: taking traditional Chinese medicine lobelia as the main subject, search the related journal literatures of CNKI and Wanfang medical database (January 2000.01 to August 2018.10), collect the prescriptions recorded in literatures, establish a database and conduct statistical analysis. Statistical software Excel 2013, SPSS Modeler 14.1 and SPSS Statistics 19.0 were used to conduct frequency analysis, association rule analysis and cluster analysis for the included traditional Chinese medicine. Result: a total of 45 journal literatures were included, containing 165 tastes of traditional Chinese medicine, with a total frequency of 406 times. Frequency analysis found that among the 45 prescriptions included in the inclusion criteria, lobelia (45 times, 100.0%), sculellaria barbata (29 times, 64.4%), oldenlandia diffusa (26 times, 57.8%), poria cocos (11 times, 24.4%) appeared most frequently. In property and flavor, most of the medicine tastes sweet (104 times, 35.99%), acrid (85 times, 29.41%) and cold (100 times, 52.08%), neutral (75 times, 39.06%) in property. In the channel tropism, lung meridians (117, 20.74%), heart meridians (107, 18.97%) and liver meridians (91, 16.13%) were the most common. There are 10 kinds of common TCM diseases, most of which are ear, throat, and oral cavity. 12 drug pairs with the highest intensity of association were obtained in association rule analysis. In the system cluster analysis, the core group square 6 was obtained. Conclusion: by applying modern information technology and combining TCM clinical data with big data for in-depth analysis and integration, it will be more convenient to explore the potential compatibility law of TCM, and finally feedback to clinical practice, providing basis for guiding the research and development of clinical medication and new prescription.