Machine learning classifiers for attributing tephra to source volcanoes: an evaluation of methods for Alaska tephras
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
ABSTRACT Glass composition‐based correlations of volcanic ash (tephra) traditionally rely on extensive manual plotting. Many previous statistical methods for testing correlations are limited by using geochemical means, masking diagnostic variability. We suggest that machine learning classifiers can expedite correlation, quickly narrowing the list of likely candidates using well‐trained models. Eruptives from Alaska's Aleutian Arc‐Alaska Peninsula and Wrangell volcanic field were used as a test environment for 11 supervised classification algorithms, trained on nearly 2000 electron probe microanalysis measurements of glass major oxides, representing 10 volcanic sources. Artificial neural networks and random forests were consistently among the top‐performing learners (accuracy and kappa > 0.96). Their combination as an average ensemble effectively improves their performance. Using this combined model on tephras from Eklutna Lake, south‐central Alaska, showed that predictions match traditional methods and can speed correlation. Although classifiers are useful tools, they should aid expert analysis, not replace it. The Eklutna Lake tephras are mostly from Redoubt Volcano. Besides tephras from known Holocene‐active sources, Holocene tephra geochemically consistent with Pleistocene Emmons Lake Volcanic Center (Dawson tephra), but from a yet unknown source, is evident. These tephras are mostly anchored by a highly resolved varved chronology and represent new important regional stratigraphic markers.
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 | 0.016 | 0.003 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.000 | 0.001 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.001 |
| Open science | 0.002 | 0.000 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.000 | 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