采用人工神经网络模型评价喜炎平辅助治疗婴幼儿急性支气管炎的疗效

Evaluation of the efficacy of Xiyanping in adjuvant treatment of acute bronchitis in infants and young children using artificial neural network model

目的:采用人工神经网络模型评价喜炎平辅助治疗婴幼儿急性支气管炎的疗效。方法:收集成都地区2017-2018年三级甲等医院门诊确诊及入院治疗的婴幼儿病例,采用人工神经网络模型modified globally convergent version(GRPROP) resilient backpropagation (RPROP)算法的建立,评价喜炎平辅助治疗婴幼儿急性支气管炎的疗效,从药效学的角度阐述其辅助用药的有效性。结果:以出院疗效显著性、喘息存在时长(止喘能力)、咳嗽存在时长(止咳能力)、鼻塞存在时长(治疗鼻塞能力)、发热存在时长(解热能力)为判定目标构建人工神经网络模型,将急性支气管炎入院的婴幼儿患者根据感染源区分为三种亚型,即生物学意义可解释为感染源,病毒、细菌或混合型。喜炎平作为辅助药物均参与三种亚型病例的治疗,对三种亚型的权重值分别为1.29611、1.23571、1.78513,说明喜炎平对三种亚型患者均能产生一定的治疗效应,且效应相当。结论:本研究构建了鲁棒性的GRPROP RPROP的喜炎平儿科患者的BP网络模型,从模型可见其对三种类型的感染性支气管炎串者均可加快好转,但作用较为温和,可提供约10%的对抗病毒药物的增效作用。由于无进行抗细菌治疗的相关病例,其与抗菌药物的相关联合研究有待进一步收集病例后加以分析。

Objective:To evaluate the efficacy of Xiyanping in adjuvant treatment of acute bronchitis in infants and young children using an artificial neural network model.Method:Infant cases diagnosed as acute bronchitis and admitted to the hospital from 2017 to 2018 in the Third-level grade-A hospitals in Chengdu were collected.The artificial neural network model modified globally convergent version(GRPROP) resilient backpropagation(RPROP) algorithm was established to evaluate the adjuvant treatment of infants with Xiyanping,to explain the effectiveness of its adjuvant medication from the perspective of pharmacodynamics.Result:The artificial neural network model was constructed based on the significance of the discharge effect,the duration of wheezing(anti-asthmatic ability),the duration of coughing(the ability to suppress cough),the duration of nasal congestion(the ability to treat nasal congestion),and the duration of fever(antipyretic ability).Infants and young children admitted to hospital with acute bronchitis were divided into three subtypes according to the source of infection,that is,the biological significance can be explained as the source of infection,virus,bacteria or mixed type.As an adjuvant drug,Xiyanping was involved in the treatment of three subtypes.The weight values for the three subtypes were 1.29611,1.23571,1.78513 respectively,indicating that Xiyanping can had a certain therapeutic effect on three subtypes with equal effect.Conclusion:This study constructs a robust GRPROP RPROP BP network model of Xiyanping in trreating pediatric patients.It can be seen from the model that it can speed up the improvement of three types of infectious bronchitis,but the effect is relatively mild and can provide about a 10% synergistic effect of antiviral drugs.Because there are no related cases of antibacterial treatment,the related joint research with antibacterial drugs needs to be further collected and analyzed after the case is collected.