Dataset and Code for “A Domain-Informed Automated Machine Learning Framework for Dielectric Ceramic Design: Application to the BaTiO₃–Bi(Mg₁/₂Ti₁/₂)O₃ System”

This dataset contains the raw data, feature descriptors, and Python source codes used in the study “A Domain-Informed Automated Machine Learning Framework for Dielectric Ceramic Design: Application to the BaTiO₃–Bi(Mg₁/₂Ti₁/₂)O₃ System”.
It includes the curated dielectric property dataset (313 entries), feature-selection scripts, Auto-Sklearn/Auto-Keras model configurations, and the uncertainty quantification module for predicting dielectric constants and X9R compliance.
The dataset supports reproducibility and reuse for research on automated machine learning, uncertainty quantification, and dielectric ceramic materials design.

本数据集包含研究论文《面向领域知识的自动化机器学习框架用于介电陶瓷设计：在钛酸钡-钛镁酸铋（BaTiO₃–Bi(Mg₁/₂Ti₁/₂)O₃）体系中的应用》中所使用的原始数据、特征描述符以及Python源代码。本数据集涵盖经整理的介电性能数据集（共313条条目）、特征选择脚本、Auto-Sklearn/Auto-Keras模型配置，以及用于预测介电常数与X9R合规性的不确定性量化（uncertainty quantification）模块。本数据集可为自动化机器学习、不确定性量化以及介电陶瓷材料设计领域的相关研究提供可复现性与复用性支持。