优化的Apriori算法应用于中风病的处方配伍规律研究

发布时间：2018-09-10 21:18

【摘要】：目的:在传统关联规则“支持度-置信度”模式的基础上引入新的参数Kulc和IR参数,以优化挖掘结果的准确性;对比分析优化前后的挖掘结果,证实新参数的引入能够更加深入有效的得到有意义的配伍规律;挖掘出的结果可以应用于现代中医的临床教学、理论研究以及与大数据、互联网+相结合的中医类应用平台等。方法:对目前中医领域的研究热点中风病病历信息建立SQL Server数据库,使用Apriori算法对其方剂配伍规律进行挖掘分析;使用Microsoft Visual Studio 2010编写程序,引入相关性度量参数Kulc和不平衡比(IR)参数,避免“伪强规则”的出现,确保挖掘出真正有强相关性的规则;结合中医理论对优化后的挖掘结果进行中医分析,确定配伍规律。结果:通过Apriori算法和新参数Kulc和IR,得到优化前后的中风病方剂配伍规律;对比优化前后的挖掘结果,发现优化后确实能够过滤掉许多相关性不强的药物组合,得到真正意义上的药对药组;中医分析结果显示挖掘出的药物确实在治疗中风病时配伍使用效果更佳,为现代中医治疗中风病提供数据支持和理论依据。结论:通过引入Kulc和IR参数确实可以过滤非强关联的配伍规则,提高了关联规则挖掘的准确度,得到了可应用的现代中医治疗中风病的配伍规律;新方法同样可应用于其它疾病的诊疗规律挖掘,尤其是用药规律尚未完全明确的疾病,可以为中医理论的学习和临床提供理论支持和教学引导。
[Abstract]:Aim: to optimize the accuracy of mining results by introducing new parameters Kulc and IR on the basis of the traditional support and confidence model of association rules, and to compare and analyze the mining results before and after optimization. The results can be applied to the clinical teaching of modern Chinese medicine, the theoretical research and the application platform of TCM combined with big data and Internet. Methods: the SQL Server database was established for the information of stroke medical records, which is a hot research topic in the field of traditional Chinese medicine, and its prescription compatibility law was excavated and analyzed by using Apriori algorithm, and the program was compiled with Microsoft Visual Studio 2010. In order to avoid the appearance of pseudo-strong rule, the correlation measure parameter Kulc and the unbalanced ratio (IR) parameter are introduced to ensure that the rules with real strong correlation can be excavated. Combined with TCM theory, the optimized mining results are analyzed in TCM, and the compatibility law is determined. Results: by using Apriori algorithm and new parameters Kulc and IR, the compatibility of stroke prescription before and after optimization was obtained, and by comparing the mining results before and after optimization, it was found that the optimized combination could filter out many drug combinations with no strong correlation. The results of traditional Chinese medicine analysis show that the medicine is better in the treatment of apoplexy, which provides data support and theoretical basis for modern Chinese medicine treatment of apoplexy. Conclusion: by introducing the parameters of Kulc and IR, the compatibility rules of non-strong association can be filtered, the accuracy of association rules mining can be improved, and the compatibility law of modern TCM treatment of apoplexy can be obtained. The new method can also be applied to excavate the rule of diagnosis and treatment of other diseases, especially the disease which has not completely defined the law of drug use, which can provide theoretical support and teaching guidance for the study of TCM theory and clinical practice.
【学位授予单位】：山东中医药大学
【学位级别】：硕士
【学位授予年份】：2016
【分类号】：TP311.13;R255.2

【参考文献】