基于决策树的卫生装备数质量预警模型的应用探讨

摘要

Objective To propose a decision tree-based early warning model so as to predict the medical equipment quantity and quality under special conditions to quantize maintenance and detection staffs allocation,components supply,emergency planning and etc.Methods A data set was established based on the verification report,data on performance parameters detection and daily management log,and then underwent pretreatment.A decision tree came into being with the algorithms of ID3,CHAID and etc.The data rule corresponding to the decision tree was transformed into a series of early warning information.The model was verified by applying it to medical equipment usage management.Results The decision tree developed was applied to analyzing a vehicle-mounted water purifying device maintenance and usage records in some logistics support vehicle,and it's found the main factors contributing to the failures included migration distance along non-paved road,ambient temperature and service time.Conclusion The decision tree-based early warning model for medical equipment quantity and quality gains high feasibility and practical values.%目的:研究基于决策树的卫生装备数质量预警模型,实现准确预测卫生装备在未来特定条件下的数质量情况,为制定决策提供技术支持,量化指导维修检测力量配置、零配件供应、应急预案制订等工作.方法:以检定(校准、检测)报告、性能参数检测数据、日常使用管理记录为数据源建立数据集,首先对数据集进行预处理,之后利用ID3、CHAID等算法生成决策树,最后将决策树反映的数据规则转换成一系列预警信息.将该模型应用于卫生装备使用实践,进行分析验证.结果:对某后勤保障车辆车载净水装置维护使用档案记录进行决策树分析发现,非铺装路面机动距离、环境温度、服役时间对故障发生率有主要影响.结论:建立基于决策树的卫生装备数质量预警模型具有一定的可行性和实用价值.