Scholarlink

<p>Machine learning (ML) has the potential to advance materials science. However, ML algorithm selection is often <i>ad hoc</i>; limiting the performance and impact of ML, and guidance on choosing appropriate algorithms is scarce. We explore a subset of ML algorithms applied to diverse applications involving both experimental and computational data, and utilizing regression, interpolation, and optimization ML methods. We study three applications involving distinct data regimes and goals, allowing us to compare the performance of algorithms. We provide practical insights into algorithm selection, frameworks for studying algorithmic choice, and encourage further investigation into this critical but underexplored aspect of ML.</p> Graphical Abstract <p></p>

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

Guidance on machine learning algorithm selection for materials science and engineering applications

  • 期刊论文

  • 发表日期：

    2025年10月1日

查看全文
  • Ziqing Zhao,
  • Emily M. Ryan

摘要

Machine learning (ML) has the potential to advance materials science. However, ML algorithm selection is often ad hoc; limiting the performance and impact of ML, and guidance on choosing appropriate algorithms is scarce. We explore a subset of ML algorithms applied to diverse applications involving both experimental and computational data, and utilizing regression, interpolation, and optimization ML methods. We study three applications involving distinct data regimes and goals, allowing us to compare the performance of algorithms. We provide practical insights into algorithm selection, frameworks for studying algorithmic choice, and encourage further investigation into this critical but underexplored aspect of ML.

Graphical Abstract