Dataset for Adaptive sampling-based surrogate modeling for composite performance prediction

Description

This dataset contains the raw data for the paper titled "Adaptive sampling-based surrogate modeling for composite performance prediction" and the trained Gausssian process regression (GPR) model from adaptive sampling. Folder Summary: The root folder contains a subfolder 'saved_model' and relevant files for training. -The `saved_model` subfolder contains: Multiple versions of serialized models (`.joblib` and `.pkl`), reflecting variations in model configurations or training epochs. -'model_20_5_20.pkl' file contains the settings of a saved model with 20 seed data, 5 queries and 20 sampling size in Pickel format -'model_20_5_20.joblib' file contains the settings of a saved model with 20 seed data, 5 queries and 20 sampling size in Joblib format - Training notebook (`AL_gp_training.ipynb`) to manage and execute training workflows. - Dataset (`Data_all.csv`) in CSV format for training and testing purposes. -Readme ('Readme.md') to run the code

Steps to reproduce

The homogenized stress-strain data are generated from abaqus simulation. For the simulation setup details, please refer to the original paper. The trained machine learning model and the code is provided in this dataset.

Categories

Funding

Licence