Reproducibility Package: Stacking Ensemble Machine Learning with Dual-Level SHAP Explainability for Precision Drip Irrigation Scheduling on Saline Clay Soils in the Nile Delta

Description

This dataset provides a focused reproducibility package for the manuscript: "Stacking Ensemble Machine Learning with Dual-Level SHAP Explainability for Precision Drip Irrigation Scheduling on Saline Clay Soils in the Nile Delta" The package contains the essential materials required to reproduce the machine learning pipeline and results reported in the study. The uploaded files include: • 01_models Trained stacking ensemble models developed for precision drip irrigation scheduling. • 02_scalers_transformers Fitted scalers and transformers are used during data preprocessing and feature scaling. • 3_environment Computational environment specifications (dependencies, library versions, and configuration details) required to recreate the modeling environment. • README and README_v2 Detailed documentation explaining the structure of the repository, how to load the models and scalers, and steps to reproduce the key results. These materials enable independent researchers to load the trained models, apply the same preprocessing transformations, and verify the reported performance under the specified computational environment. The package supports full transparency and reproducibility of the stacking ensemble approach and dual-level SHAP explainability analysis presented in the manuscript.

Steps to reproduce

1. Download and extract all files from this dataset. 2. Set up the computational environment using the specifications provided in the "3_environment" folder (install the required Python packages and versions). 3. Load the trained stacking ensemble models from the "01_models" folder. 4. Apply the fitted scalers and transformers from the "02_scalers_transformers" folder to preprocess new or test data using the same pipeline described in the manuscript. 5. Generate predictions and compare the results with the performance metrics reported in the paper. 6. Refer to the README and README_v2 files for detailed instructions on loading the models and reproducing the dual-level SHAP explainability analysis.

Institutions

• Tarim University

  Xinjiang, Aral

Categories

Licence