基于机器学习的多参数监护仪维护需求预测分析方法
The Predictive Analysis Method for Maintenance Demand of Multi-Parameter Monitor Based on Machine Learning
李坤 1杨秉泽 1穆骞1
作者信息
- 1. 三二〇一医院(陕西汉中 723000)
- 折叠
摘要
目的 利用机器学习(ML)算法分析多参数监护仪质控数据集建立预测系统,并探讨其对下个质控周期多参数监护仪维护需求的预测效果.方法 筛选 2020 年 1 月至2023 年 12 月医院 1 500 条多参数监护仪质控数据作为原始数据集,并按照8‥2比例划分为训练集和测试集,训练集数据1 200条,测试集数据300条,对训练集数据集进行特征选择,生成4组子数据集.应用朴素贝叶斯(NB)、决策树(DT)、随机森林(RF)、k-最近邻(k-NN)和支持向量机(SVM)5 种ML算法建立预测系统,并对下个质控周期多参数监护仪维护需求进行预测.结果 以实际需要进行维护的多参数监护仪为准,5 种ML算法预测多参数监护仪维护需求的平均精准率为 96.73%,真阴性率为 98.00%.结论 应用ML算法可以有效预测多参数监护仪在下个质控周期的维护需求,为多参数监护仪的维护管理提供新方案.
Abstract
Objective To analyze the quality control data set of multi-parameter monitor by using machine learning(ML)algorithm to establish a prediction system,and to explore the prediction effect of the maintenance demand of multi-parameter monitor in the next quality control cycle.Methods With the 1500 multi-parameter monitor quality control data from hospitals from January 2020 to December 2023 screened as the original data set,these data were divided into training and testing sets in an 8:2 ratio,with 1200 training sets and 300 testing sets.The feature selection was carried out for the training set data set,and four groups of sub-datasets were generated.With the application of five ML algorithms:naive bayes(NB),decision tree(DT),random forest(RF),k-nearest neighbor(k-NN)and support vector machine(SVM)to establish a prediction system,the prediction of maintenance needs for multi-parameter monitors in the next quality control cycle was carried out.Results Based on the actual number of multi-parameter monitors that need to be maintained,the average accuracy of the five ML algorithms in predicting the need to maintain the multi-parameter monitor was 96.73%,with a true negative rate of 98.00%.Conclusion The application of ML algorithm can effectively predict the maintenance requirements of multi-parameter monitors in the next quality control cycle,providing a new solution for the maintenance management of multi-parameter monitors.
关键词
机器学习/质控周期/多参数监护仪/维护管理Key words
Machine learning/Quality control cycle/Multi-parameter monitor/Maintenance management引用本文复制引用
出版年
2024
NETL
NSTL
万方数据