关联规则在中医肝病电子病历数据分析中的应用研究

Abstract

目的：基于关联规则挖掘模型分析中医肝病患者数据，探寻检查指标与中医辩证之间的关联。方法：通过设置最小支持度和最小可信度对中医肝病患者资料进行关联规则分析，根据生成规则的重要性筛选出前后件成正相关的规则，结合提升图评价挖掘结果准确性。结果：分析样本例数317例，共获得30条规则，揭示了检查指标组合与中医辨证结果间的关系。结论：在中医肝病电子病历数据中应用关联规则分析可以揭示不同检查指标对于中医辨证的影响，有利于辅助诊断。 Objective: Based on analysis association rules mining model data for the liver disease of traditional Chinese medicine (TCM), to search for relation between checking index and the Chinese medicine dialectical. Method: By setting the minimum support and minimum confidence, we used association rules to analyze the patients’ data for the liver disease of traditional Chinese medicine. According to the importance of generated rules, we screened out the rules which are positively correlated to the before and after rule, and evaluated the accuracy of the mining results with the lift chart. Result: Through the analysis of 317 samples, 48 rules are received, which reveals the relationship between the check index combination and TCM syndrome differentiation results. Conclusion: Using association rules in electronic medical record data analysis for the liver disease of TCM can reveal the influence of different examination indexes for TCM syndrome differentiation, and is advantageous to the auxiliary diagnosis.