首页> 外文会议>Computational methods in science and engineering: advances in computational science >Kaolin Quality Prediction from Samples: A BayesianNetwork Approach
【24h】

Kaolin Quality Prediction from Samples: A BayesianNetwork Approach

机译:样品的高岭土质量预测:贝叶斯 r n网络方法

获取原文
获取原文并翻译 | 示例

摘要

We describe the results of an expert system applied to the evaluation of samples of kaolin for industrial use in paper or ceramic manufacture. Different machine learning techniques—classification trees, support vector machines and Bayesian networks—were applied with the aim of evaluating and comparing their interpretability and prediction capacities. The predictive capacity of these models for the samples analyzed was highly satisfactory, both for ceramic quality and paper quality. However, Bayesian networks generally proved to be the most useful technique for our study, as this approach combines good predictive capacity with excellent interpretability of the kaolin quality structure, as it graphically represents relationships between variables and facilitates what-if analyses.
机译:我们描述了专家系统的结果,该专家系统用于评估造纸或陶瓷制造工业用高岭土样品。为了评估和比较它们的可解释性和预测能力,应用了不同的机器学习技术(分类树,支持向量机和贝叶斯网络)。这些模型对所分析样品的预测能力在陶瓷质量和纸张质量方面都非常令人满意。然而,贝叶斯网络通常被证明是对我们的研究最有用的技术,因为这种方法结合了良好的预测能力和高岭土质量结构的出色可解释性,因为它以图形方式表示变量之间的关系并有助于进行假设分析。

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱:kefu@zhangqiaokeyan.com

京公网安备:11010802029741号 ICP备案号:京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有
  • 客服微信

  • 服务号