基于机器学习的非法使用精神、麻醉药品行为模型的研究  

Research on the Behavior Model of Illegal Use of Psychotropic Substances and Narcotic Drugs Based on Machine Learning

作　　者：陈娉娉 CHEN Ping-ping(Xiamen Health and Medical Big Data Center,Fujian Xiamen 361008)

机构地区：[1]厦门市健康医疗大数据中心,福建厦门361008

出　　处：《中国医疗器械信息》2025年第6期156-158,共3页China Medical Device Information

基　　金：福建省卫健委科技计划项目(项目名称:厦门市医护人员特定医疗行为监管分析,项目编号:2020QNB070)。

摘　　要：目的:探索基于机器学习算法预测是否非法使用精神、麻醉药品的行为,以达到智能监管、识别及预警的目的。方法:将27109份使用精神麻醉药品人群的全样本诊疗数据进行清洗和特征工程处理,利用Python进行模型的建模效果评估及检验,对模型进行训练、测试,采用多种算法进行比对,选出最优模型。结果:随机森林构建的模型预测结果最好,模型整体准确率为0.99。结论:基于随机森林模型构建的非法使用精神、麻醉药品模型整体正确率较高,基本能够满足用药预测预警分析需求,具有一定的实际应用价值,为利用智能化提高卫生监督管理水平提供了思路。Objective:Exploring prediction model that identification illegal behavior about using psychotropic and narcotic drugs based on machine learning algorithm,in order to achieve the purpose of intelligent supervision,identification and early warning.Methods:27109 samples data of people using psychotropic drugs were cleaned and processed by feature engineering.As to select the best models,python is used to evaluate and test the model,and the model was trained and tested.A variety of algorithms were used to compare and select the optimal model.Results:The model constructed by random forest get the best prediction result,and the overall accuracy of the random forest model was 0.99.Conclusion:The model based on the random forest model has the highest accuracy on illegal behavior prediction of using spirit and narcotic drugs,which can basically meet the needs of drug use prediction and early warning analysis.It has a certain practical value,and provides a way of thinking for improving the level of health supervision and management by using intelligence.

关 键 词：卫生监管 随机森林 预测模型 

分 类 号：R197.39[医药卫生—卫生事业管理]