目的 选用两种药物分析对血脂异常、中医证候的疗效，以中医证候积分为基础建立血脂异常气郁痰阻型的疗效表征模型，进而提出基于粒子群优化的BP神经网络预测两种方法的疗效。方法 收集50名符合血脂异常气郁痰阻型患者的基本资料，将患者随机分为解郁祛痰化浊方组（JQHP）、血脂康胶囊组（XZKC），每组25例，观察疗程12周治疗前后的血脂水平，治疗前及治疗期间每4周记录一次中医证候积分，共4次。定性疗效分析参考血脂防治指南和中医疗效积分，采用t检验、非参数检验及X²分析的方法；神经网络模型采用粒子群优化的BP神经网络，定量构建血脂异常气郁痰阻证的预测模型，设置为三层网络，输入层节点数为15，隐含层节点数为6，输出层节点数为1。针对JQHP组和XZKC组的数据，分别选择80%的数据进行训练，20%的数据进行验证。结果 两组治疗前、治疗4周、治疗8周、治疗12周的中医证候积分随治疗周期的增加而下降（P < 0.05）。JQHP组除了“形体肥胖”外，其余14项症状均有显著的改善（P < 0.05），XZKC组治疗前后比较除“形体肥胖”、“头晕”、“胸闷”外，其余12项均改善（P < 0.05）。以中医证候积分建立的基于粒子群优化BP神经网络的预测精度高、稳定性佳，可以定量预测两种药物对未知患者的疗效。结论 JQHP和XZKC可以改善血脂异常患者的血脂水平，改善气郁痰阻型患者的症状，此为定性分析。利用中医证候积分建立的基于粒子群优化BP神经网络两种治疗方法的预测模型，可定量预测未知患者的疗效。以此思路将有助于找到一种特定治疗方法的适应人群，从而提高该治疗方法的针对性，实现中医药的智能化、精准化、定量化。

Objective Two drugs were selected to analyze the efficacy of Dyslipidemia and TCM syndromes. Based on the TCM syndrome scores, the efficacy characterization model of Qi Stagnation and Phlegm Obstruction Syndrome in Dyslipidemia were established, and then BP Neural Network based on Particle Swarm Optimization were proposed to predict the therapeutic efficacy of the two methods.Methods The basic data of 50 patients on dyslipidemia of Qi Stagnation and Phlegm Obstruction Syndrome were collected. The patients were randomly divided into Jieyu Qutan Huazhuo Prescription group (JQHP) and Xuezhikang Capsule group (XZKC), with 25 cases in each group. The blood lipid level was observed for 12 weeks before and after treatment. The TCM syndrome scores were recorded every 4 weeks before and during treatment, a total of 4 times. Qualitative efficacy analysis refers to the guidelines for prevention and treatment of blood lipids and TCM syndrome scores, using t-test, nonparametric test and X2 analysis. The BP neural network based on Particle Swarm Optimization to quantitatively construct the prediction model on dyslipidemia of qi stagnation and phlegm obstruction syndrome. It is set as a three-layer network, with 15 nodes in the input layer, 6 nodes in the hidden layer and 1 node in the output layer. For the data of JQHP group and XZKC group, 80% of the data were selected for training and 20% of the data were verified.Results The TCM syndrome scores of the two groups before treatment, after treatment 4 weeks, 8 weeks and 12 weeks decreased with the increase of treatment time (P < 0.05). In JQHP group, except for "physical obesity", the other 14 symptoms were significantly improved (P < 0.05). In XZKC group, except for "physical obesity", "dizziness" and "chest tightness", the other 12 symptoms were improved (P < 0.05). The particle swarm optimization BP neural network based on TCM syndrome scores have high prediction accuracy and good stability. It can quantitatively predict the curative effect of two drugs on unknown patients.Conclusion JQHP and XZKC can improve the blood lipid level of patients with dyslipidemia and improve the symptoms of patients with qi stagnation and phlegm obstruction. This is a qualitative analysis. The prediction models of two treatment methods based on BP Neural Network Optimized by Particle Swarm established by TCM syndrome scores can quantitatively predict the curative effect of unknown patients. This idea will help to find an adaptive population for a specific treatment method, so as to improve the pertinence of the treatment method and realize the intelligence, accuracy and quantification of TCM.

R2-03

中国中医科学院中医药“一带一路”合作专项(GH201911):基于luciferase reporter实验技术联合实验室研发改善血脂代谢异常的中草药、美洲药用植物,负责人:姚魁武；中国中医科学院自主选题第十一批(ZZ11-031):血脂异常社会学病因分析及解郁祛痰化浊方干预研究,负责人:姚魁武