支持向量机模型在有机溶剂职业危害评价中的应用研究  

Application of support vector machines on the prediction of occupational health Assessment

作　　者：温翠菊[1] 李荣宗[1] 黎丽春[1] 温薇[1] 陈建雄[1] 

机构地区：[1]广东省职业病防治院职业卫生评价所,广东广州510300

出　　处：《中国卫生工程学》2015年第1期23-26,共4页Chinese Journal of Public Health Engineering

基　　金：广东省医学科学技术研究基金(B2011021)

摘　　要：目的初步探讨支持向量机(support vector machines,简称SVMs)模型在有机溶剂职业危害评价中的应用。方法选择17家使用有机溶剂企业,对其使用有机溶剂作业进行调查,根据现场检测、职业卫生学调查、劳动者职业健康检查等资料,筛选构建评价模型指标体系。运用SVMs方法,采用R软件的e1071软件包,对17家企业14种有机溶剂作业职业危害进行训练和预测。结果当选用10项评价指标时,预测准确率为85.45%,当选用8项评价指标时,预测准确率为69.09%,当选用4项评价指标时,预测准确率为80.00%。结论 SVMs模型可以对有机溶剂职业危害进行预测,对建设项目职业危害评价工作具有一定参考意义,但其实际应用需进一步验证。Objective To explore the application of support vector machines（ SVMs） assessment methods on occupational health assessment. Methods Fourteen kinds of solvents were included in this survey from 15 different industries. Exposure level of these solvents,occupational hygienic survey and annual physical exam data were all included in this study. SVMs method was applied using e1071 package in R software and predict the control effect of each chemical in each factory. Results The accuracy of prediction was 85. 45% when using 10 assessment indexes,and it was 69. 09% when using 8 indexes,while it came to be 80. 00% when using 4 indexes. Conclusion These applications suggest that SVMs methods can be used for the prediction of occupational health assessment.

关 键 词：支持向量机 有机溶剂 职业危害评价 

分 类 号：R135[医药卫生—劳动卫生]