基于置信规则库的慢性萎缩性胃炎陈永灿辨证经验挖掘

CHEN Yongcan’s Syndrome Differentiation Experience Mining for Chronic Atrophic Gastritis Based on Belief Rule Base

[目的]根据名老中医陈永灿(下称陈氏)诊治慢性萎缩性胃炎(chronic atrophic gastritis,CAG)的临床资料和诊疗方案,挖掘基于置信规则库(belief rule base,BRB)的CAG辨证经验。[方法]对症候信息进行编码,将症候"有、无"分别赋值为"1、0",以此建立CAG中医症候信息数据库。将编码后的信息作为CAG辅助诊断模型的输入特征,通过BRB挖掘症候信息与证素之间的关系,利用证据推理(evidence reasoning,ER)融合算法对输入特征进行激活处理并输出结果,最后根据输出的证素结果结合临床实际得到证型。[结果]在收集到的572例CAG病例数据集上对CAG辅助诊断模型进行训练及测试,证实该模型准确率达到79.07%,优于传统的支持向量机(support vector machine,SVM)算法;由预测结果得到的证型主要为脾虚气滞型、寒热错杂型、肝胃郁热型。[结论]CAG辅助诊断模型在实际临床运用中具有可行性,表明人工智能技术有助于名老中医经验的传承和传播。

[Objective] To mine the chronic atrophic gastritis(CAG) syndrome differentiation experience based on belief rule base(BRB), according to the clinical data and diagnosis and treatment plan of CAG by CHEN Yongcan(hereinafter referred to as CHEN ’s), a well-known traditional Chinese medicine(TCM) doctor. [Methods] The symptom information was encoded, and the "yes" and "no" of symptom information were assigned "1" and "0"respectively, so as to establish the TCM symptom information database of CAG, and the encoded information was used as the input feature of CAG auxiliary diagnosis model, the relationship between the symptom information and the evidence elements was mined through BRB, and the input feature was activated by evidence reasoning(ER) fusion algorithm, and the result was output. Finally, according to the output of the syndrome element results and the clinical reality, the syndrome type was obtained. [Results] The CAG auxiliary diagnosis model was trained and tested on the data set of 572 CAG cases. It was confirmed that the accuracy of the model was 79.07%, which was better than the traditional support vector machine(SVM) method;the syndrome types obtained from the prediction results were mainly of spleen deficiency and Qi stagnation type, cold and heat complex type, liver and stomach heat stagnation type. [Conclusion] The CAG assisted diagnosis model is feasible in actual clinical application, which shows that the artificial intelligence technology is helpful to the inheritance and dissemination of the experience of famous and old Chinese medicine practitioners.