Supporting Data for "Intelligent prediction of steel corrosion in cementitious materials via machine learning"

This dataset is associated with the PhD thesis "Intelligent prediction of steel corrosion in cementitious materials via machine learning", which focuses on the development of data-driven and physics-informed machine learning models for predicting the corrosion behavior of steel in cementitious environments. The dataset is structured according to the thesis chapters (Chapter 3 to Chapter 7), with each part containing the original experimental data and relevant code used in the corresponding analyses.

• Chapter 3 contains laboratory corrosion data of steel under carbonation conditions. The dataset includes 16 variables and 180 groups, along with code implementing relevant regression algorithms.
• Chapter 4 contains laboratory corrosion data of steel under chloride ingress conditions, comprising 15 variables and 95 groups. It also includes literature-sourced corrosion data with 5 variables and 81 groups. The folder provides code for both regression and transfer learning models.
• Chapter 5 provides data for corrosion probability prediction, including 4 variables and 535 groups. It also contains code for probabilistic classification and corrosion mapping.
• Chapter 6 includes corrosion data of steel under drying-wetting cycling conditions, with 10 variables and 284 groups. The folder also contains code for regression analysis.
• Chapter 7 provides code related to symbolic learning for interpretable corrosion modeling, based on the data compiled from previous chapters.