考虑个体差异的酒后血液乙醇浓度推测模型  

Prediction model of blood ethanol concentration after alcohol consumption considering individual differences

作　　者：肖阳 陈旭 刘延 Xiao Yang;Chen Xu;Liu Yan(School of Vehicle Engineering,Chongqing University of Technology,Chongqing 400054;China Automotive Engineering Research Institute Corporation,Chongqing 401147)

机构地区：[1]重庆理工大学车辆工程学院,重庆400054 [2]中国汽车工程研究院股份有限公司,重庆401147

出　　处：《中国法医学杂志》2023年第3期268-272,共5页Chinese Journal of Forensic Medicine

摘　　要：目的为反应真实的酒后人体血液酒精浓度(BAC)随时间的变化关系。方法本研究选取10名志愿者参加饮酒测试并采集数据,运用药物代谢动力学的方法建立了预测模型。用偏最小二乘法推导出体重、性别、酒类、饮酒量对人体酒精吸收速率的影响权重,并采用交叉验证的方法判断结果的可靠性。最后使用MATLAB软件进行拟合。结果模型的平均预测准确度达到了92.17%,与传统的BAC模型相比能更准确预测酒后人体血液酒精浓度变化。其次对数据不做任何分布假设,运用BP神经网络方法分析数据并对人体BAC进行预测,其准确度为84.62%。结论在样本量较少的情况下,本研究提出的模型预测效果更优。Objective To reflect the real relationship between human blood alcohol concentration(BAC)and time after drinking alcohol.Methods 10 volunteers were selected to participate in the drinking test and the data were collected.,The predictive model was established by the method of pharmacokinetics.The weights of the effects of body weight,gender,alcohol type,and alcohol consumption on the rate of alcohol absorption in humans were derived using partial least squares,and the reliability of the results was judged by cross-validation method.Finally,MATLAB software was used for fitting data.Results the average prediction accuracy of the model reached 92.17%,which was more accurate than the traditional BAC model.Secondly,the BP neural network method was applied to predict the human BAC without any distribution assumptions,and its accuracy was 84.62%.Conclusion the model proposed in this paper has a better prediction effect when the sample size is small.

关 键 词：个体差异 药物动力学 偏最小二乘法 人体酒精吸收速率 

分 类 号：D918.93[政治法律—法学]