A systematic review of machine learning models for groundwater level prediction
Bibliographic record
Abstract
This study presents a comprehensive synthesis of machine learning (ML) techniques applied to groundwater level (GWL) prediction, focusing on model architectures, feature selection methods, hyperparameter tuning, optimization algorithms, and clustering techniques. A total of 223 peer-reviewed articles were systematically reviewed using the PRISMA framework to guide study identification, inclusion, and exclusion. Widely used models include artificial neural networks (ANN), support vector machines (SVM), long short-term memory networks (LSTM), and random forests (RF). More recent studies increasingly employ hybrid approaches that integrate wavelet transforms, signal decomposition, and optimization techniques such as particle swarm optimization (PSO), genetic algorithms (GA), and ant colony optimization (ACO). Transformer-based models have also begun to emerge as promising tools in this domain. A central focus of this review is feature selection, which remains one of the most underdeveloped areas in GWL modeling. Most studies rely on simple filter methods like autocorrelation and mutual information. While SHapley Additive exPlanations (SHAP) has gained some traction, more advanced techniques, such as recursive feature elimination (RFE), forward feature selection (FFS), factor analysis (FA), and self-organizing maps (SOM), are rarely used. Notably, no study systematically compared multiple feature selection strategies, limiting insights into their impact on model performance. Scientometric analysis shows that Iran, China, India, and the United States contribute the most impactful research. Despite strong predictive outcomes, trial-and-error remains the dominant approach to hyperparameter tuning. The review emphasizes the need for more systematic, interpretable, and generalizable ML approaches to support robust groundwater level (GWL) forecasting.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.008 | 0.040 |
| Meta-epidemiology (narrow) | 0.002 | 0.001 |
| Meta-epidemiology (broad) | 0.004 | 0.005 |
| Bibliometrics | 0.009 | 0.009 |
| Science and technology studies | 0.000 | 0.001 |
| Scholarly communication | 0.002 | 0.002 |
| Open science | 0.002 | 0.001 |
| Research integrity | 0.002 | 0.001 |
| Insufficient payload (model declined to judge) | 0.005 | 0.001 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".