Red Sandstone Dynamic Compressive Strength Data for Wetting-Drying Cycle Prediction

A dataset of red sandstone dynamic compressive strength (DCS) under acidic wetting-drying conditions, used to train ensemble learning models. The data was authored by Bin Du and last updated on May 14, 2026. It includes input features analyzed for correlation and importance via SHAP values.

Use Cases

• Predict dynamic compressive strength of sandstone based on input features like SR, pH, Vp, n, and UCS.
• Compare performance of ensemble learning models (RF, AdaBoost, XGBoost, LightGBM, GBDT) for material property regression.
• Interpret feature importance and model predictions using SHAP analysis for local and global explanations.
• Analyze the influence of acidic wetting-drying cycles on rock strength degradation.

Strengths

• Includes model performance comparisons and error distributions for five ensemble learning algorithms.
• Provides model interpretability through SHAP summary plots, waterfall plots, and feature importance analysis.
• Released under a permissive CC-BY-4.0 license.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment for large-scale modeling.
• The 6.2 MB file size suggests a relatively small dataset.

Provenance

Source

Bin Du via figshare.

Collection Method

Likely contains experimental data from dynamic impact testing of red sandstone specimens.

Freshness

Last updated 2026-05-14 17:26:48; freshness should be verified.