中国安全科学学报

• 职业卫生工程 • 上一篇    下一篇

尘肺危害的神经网络评价及预测研究

柳静献,刘铁民,王金波   

  1. 东北大学%国家经贸委安全科学技术研究中心
  • 出版日期:2001-02-20 发布日期:2001-02-25
  • 基金资助:
    科技部科技攻关项目(NO:969180301)

Assessment and Prediction of Pneumoconiosis Hazard with Neural Network

  • Online:2001-02-20 Published:2001-02-25

摘要: 尘肺发病率与其所受影响因素—粉尘毒性、粉尘浓度及接尘时间等之间的关系,呈非线性。笔者引入神经网络,建立了从影响因素到发病率的算法模型,并开发了相应的软件(PHAF)。该软件在学习样本数据基础上,可以对尘肺危害进行评价和预测。

Abstract: The correlation of the incidence of pneumoconiosis with dust toxicity, dust concentration, and exposure duration is non-linear. In this paper, neural network has been applied to build a model for the incidence of pneumoconiosis and the influencing factors, and to develop a special software. By learning from samples, this software could be used to assess and predict accurately the pneumoconiosis hazard.

中图分类号: