Human-based metaheuristics and non-parametric learning for groundwater-prone area mapping
Effective Groundwater Potential Map (GPM) is crucial for sustainable water resource management, particularly in semi-arid areas. Existing GPM techniques often depend on parametric models, which may fail to capture the intricate patterns of groundwater distribution or adapt to varying data complexiti...
Saved in:
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Elsevier
2025-12-01
|
| Series: | Ecological Informatics |
| Subjects: | |
| Online Access: | http://www.sciencedirect.com/science/article/pii/S1574954125003358 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Effective Groundwater Potential Map (GPM) is crucial for sustainable water resource management, particularly in semi-arid areas. Existing GPM techniques often depend on parametric models, which may fail to capture the intricate patterns of groundwater distribution or adapt to varying data complexities. Traditional methods often struggle with optimizing model performance and handling complex, nonlinear relationships in environmental factors. This study addresses these challenges by integrating human-inspired metaheuristics with non-parametric machine-learning techniques to enhance groundwater potential prediction. This research introduces a novel approach combining human-based metaheuristics—Teaching Learning Based Optimization (TLBO) and Cultural Algorithms (CA)—with non-parametric Decision Tree (DT) models. We leverage TLBO and CA for hyperparameter tuning, optimizing model performance in predicting groundwater potential. The findings indicate that the DT-TLBO attained superior performance, achieving an Area Under the Curve (AUC) value of 96.5 %, surpassing the DT-CA at 94.9 % and the standalone DT at 91 %. Validation using Friedman and Wilcoxon signed-rank tests confirmed the statistical significance of our model improvements. The DT-TLBO model demonstrated superior accuracy and reliability, making it a promising tool for groundwater resource assessment. Feature importance analysis using DT-TLBO identified elevation, rainfall, and Topographic Wetness Index (TWI) as the most influential factors in GPM. This research underscores the effectiveness of integrating human-based metaheuristics with non-parametric learning to improve predictive modeling in environmental applications. |
|---|---|
| ISSN: | 1574-9541 |