Scholarlink

In this chapter, we will use the RML sampling strategy introduced in the previous chapter to introduce another simplification that eliminates the need for adjoint models. We will show that we can approximate the model sensitivityModelsensitivity by a linear regression expressed by covariance matrices. Thus, we replace the adjoint and tangent linear models with a least squares best-fitLeast squares best fit linear regressionLinear regression approximation.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Averaged Model Sensitivity

  • 章节

  • 出版模式：

    开放获取

  • 发表日期：

    2026年

查看全文
  • Geir Evensen,
  • Dean S. Oliver,
  • Remus G. Hanea

摘要

In this chapter, we will use the RML sampling strategy introduced in the previous chapter to introduce another simplification that eliminates the need for adjoint models. We will show that we can approximate the model sensitivityModelsensitivity by a linear regression expressed by covariance matrices. Thus, we replace the adjoint and tangent linear models with a least squares best-fitLeast squares best fit linear regressionLinear regression approximation.