Fuzzy-based input method for uncertainty quantification in a deterministic model comparison with ChatGPT for peak flow prediction
Bibliographic record
Abstract
• Compared model performance of PCSWMM and GPT models for peak-flow prediction under varying data-split approaches using comprehensive evaluation metrics. • Developed a novel integration of fuzzy methods for quantifying uncertainty in rainfall and peak-flow observations. • Quantified the uncertainty in model prediction (peak flow), enhancing real-world plausibility and interpretability of predictions, and effectively conveying this uncertainty to stakeholders for informed decision-making. ChatGPT, a generative AI, is applied and compared to the PCSWMM hydrological model for modelling peak flow in a small watershed in the runoff period of April to September. A new approach for fuzzy mathematical representation of rainfall and peak-flow errors was developed to lead to a fuzzy based GPT model and fuzzy based PCSWMM model. This led to fuzzy output for both models and a more appropriate application of both models given data errors and large language model structure. Training and validation were conducted with an approximately 25/75 split of the data and again using a 75/25 data split. Evaluation metrics were used to compare model performance under the different data-split scenarios. Calibrated and validated PCSWMM outperformed GPT in the 25/75 data split but ChatGPT 4o mini’s generation outperformed PCSWMM in the 75/25 split and with comparable validation metrics and an application that was less onerous than when using PCSWMM. The fuzzy-based error analysis showed that for both models, a fuzzy-based approach produced more interpretable and reasonable results than either original model. Moreover, the trade-off between coverage (uncertainty range) and precision for GPT‑4o mini model’s fuzzy output at high membership levels demonstrated enhanced predictive performance under data‑scarce conditions.
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 | 0.002 | 0.005 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.001 | 0.001 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.001 | 0.001 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.002 | 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 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".