Bibliographic record
Abstract
I assessed the performance characteristics of the feed-forward artificial neural network (ANN) as a first-order nonlinear Markov modelling technique. The ability to recover the underlying structure of five synthetic random time series was first tested. The method was then applied to an observed geophysical time series, and the results were compared against external empirical constraints and a simple representation of the underlying physics. The Monte Carlo experiments suggested that the ANN–Markov technique: (i) yields good prediction skill; (ii) in general, accurately retrieves the form of the iterative mapping, even for extremely noisy data; (iii) accomplishes the foregoing without any need to consider or adjust for the distributional characteristics of the data or driving noise; and (iv) accurately estimates the distribution of the strictly stochastic signal component. Application to a historical river-flow record again showed good forecast skill. Moreover, the robustness, flexibility, and simplicity of the method permitted easy identification of the fundamental nonlinear physical dynamics of this environmental system directly from the time series data, perhaps belying the common perception of ANNs as a strictly black-box prediction technique. The ANN–Markov technique may thus serve as a valuable data-driven tool for guiding the development of both process-based and parameteric statistical models. The lack of specific distributional assumptions and requirements notwithstanding, it was also found that manual distributional transformations may permit the method to be tuned to particular applications by emphasizing or de-emphasizing certain features of the data. Drawbacks to the method include substantial data-set length requirements, a general limitation of ANNs, as well as an inconsistent but potentially troubling tendency to partially imprint the form of the ANN activation function upon the estimated recursion relationship. PACS Nos.: 02.50.Ga, 05.10.–a, 05.45.Tp, 07.05.Mh, 02.50.Ey, 92.40.Fb
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 | 0.008 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.001 | 0.001 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.001 | 0.001 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.001 | 0.001 |
| 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 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".