Enhancing Pan evaporation predictions: Accuracy and uncertainty in hybrid machine learning models
Why this work is in the frame
A frame that forgets how it found something cannot be audited. These are the routes that admitted this work.
Bibliographic record
Abstract
Pan Evaporation (E p ) plays a pivotal role in water resource management, particularly in arid and semi-arid regions. This study assesses the predictive performance of a comprehensive range of advanced machine learning (ML) and deep learning (DL) algorithms for E p prediction using readily available environmental sensing data. The models investigated include M5 Prime (M5P), M5Rule (M5R), Kstar, their hybridized versions employing Bagging (BA), the adaptive neuro-fuzzy inference system (ANFIS), ANFIS-GA (genetic algorithm), and long short-term memory (LSTM) networks. A 30-year dataset of monthly meteorological observations (1988–2018) from the Kermanshah synoptic station in Iran served as the basis for this analysis, incorporating variables such as temperature, relative humidity, solar exposure, wind speed, and rainfall. Eight input scenarios were developed using both manual and automated feature selection techniques, including correlation-based subset selection evaluation (CfsSubsetEval or simply CSE), Principal Component Analysis (PCA), and the Relief Attribute Evaluator (RAE). The results demonstrate that the BA-Kstar ensemble model achieved superior performance (R 2 = 0.91, RMSE = 1.60, NSE = 0.91, and RSR = 0.30). Notably, manually constructed input scenarios outperformed automated feature selection methods, with maximum temperature emerging as the most significant predictor of E p variability. This study underscores the reliability and efficacy of hybrid ML models for E p forecasting, with significant implications for their broader application in diverse climates and geographical regions. • E p prediction by tree-, rule-, lazy- and neuron-based models is investigated. • Predictive performance of Decision Tree, ANFIS, LSTM and their ensemble models are examined. • Optimal input scenarios were investigated using feature selection and manual approaches. • BA-Kstar ensemble model exhibited superior performance in E p prediction. • The uncertainty analysis highlights that input selection impacts E p more than model selection.
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.
Full frame distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.001 | 0.000 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.000 | 0.000 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.001 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.001 | 0.000 |
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 it