首页> 外国专利> WORKFLOW-BASED VIBRATION SPECTRUM ANALYSIS MODEL OPTIMIZATION METHOD

WORKFLOW-BASED VIBRATION SPECTRUM ANALYSIS MODEL OPTIMIZATION METHOD

机译:基于工作流的振动光谱分析模型优化方法

摘要

Disclosed is a workflow-based vibration spectrum analysis model optimization method. The method comprises: initializing and determining evaluation parameters of a vibration spectrum analysis model and an optimization object of the vibration spectrum analysis model, and carrying out permutation and combination on preprocessing methods and multivariate analysis methods to obtain method combinations; calculating hyper-parameters within the various method combinations and corresponding hyper-parameter space combinations; respectively inputting a training set into the various method combinations, and carrying out optimal calculation to determine optimal hyper-parameters of the method combinations; respectively inputting the training set for training to obtain model parameters so as to acquire various combination models; and inputting a test set into the various combination models, calculating evaluation parameters of the various combination models, and selecting an optimal model. According to the present invention, a workflow is established, avoiding tedious manual operation and subjective judgment, making full use of parallel computing resources, and providing a new system model optimization method that is not available in traditional spectrum analysis software.
机译:公开了一种基于工作流的振动频谱分析模型优化方法。该方法包括:初始化并确定振动频谱分析模型的评估参数和振动频谱分析模型的优化对象,并对预处理方法和多元分析方法进行置换和组合以获得方法组合;在各种方法组合和相应的超参数空间组合内计算超参数;分别将训练集输入到各种方法组合中,并进行最优计算,确定方法组合的最优超参数;分别输入训练集进行训练以获得模型参数,以获取各种组合模型;将测试集输入到各种组合模型中,计算各种组合模型的评估参数,并选择最佳模型。根据本发明,建立了工作流程,避免了繁琐的人工操作和主观判断,充分利用了并行计算资源,并提供了一种新的系统模型优化方法,该方法在传统频谱分析软件中是不可用的。

著录项

  • 公开/公告号WO2020029851A1

    专利类型

  • 公开/公告日2020-02-13

    原文格式PDF

  • 申请/专利权人 ZHEJIANG UNIVERSITY;

    申请/专利号WO2019CN98673

  • 发明设计人 LIN TAO;XU JINFAN;YING YIBIN;

    申请日2019-07-31

  • 分类号G06F17/50;G06K9/62;G01N21/65;G01N21/3563;G01N21/359;

  • 国家 WO

  • 入库时间 2022-08-21 11:13:30

相似文献

  • 专利
  • 外文文献
  • 中文文献
获取专利

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