Development and Application of the Branched and Isoprenoid GDGT Machine Learning Classification Algorithm (BIGMaC) for Paleoenvironmental Reconstruction
Bibliographic record
Abstract
Abstract Glycerol dialkyl glycerol tetraethers (GDGTs), both archaeal isoprenoid GDGTs (isoGDGTs) and bacterial branched GDGTs (brGDGTs), have been used in paleoclimate studies to reconstruct environmental conditions. Since GDGTs are produced in many types of environments, their relative abundances also depend on the depositional setting. This suggests that the distribution of GDGTs also preserves useful information that can be used more broadly to infer these depositional environments in the geological past. Here, we combined existing iso‐ and brGDGT relative abundance data with newly analyzed samples to generate a database of 1,153 samples from several modern sedimentary settings. We observed a robust relationship between the depositional environment and the relative abundances of GDGTs in our samples. This data set was used to train and test the B ranched and isoGDGT Ma chine learning C lassification (BIGMaC) algorithm, which identifies the environment a sample comes from based on the distribution of GDGTs with high precision and recall ( F 1 = 0.95). We tested the model on the sedimentary record from the Giraffe kimberlite pipe, an Eocene maar in subantarctic Canada, and found that the BIGMaC reconstruction agrees with independent stratigraphic and palynological information, provides new information about the paleoenvironment of this site, and helps improve its paleotemperature reconstruction. In contrast, we also include an example from the PETM‐aged Cobham lignite as a cautionary example that illustrates the limitations of the algorithm. We propose that in cases where paleoenvironments are unknown or are changing, BIGMaC can be applied in concert with other proxies to generate more refined paleoclimate records.
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 | 0.004 | 0.007 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.004 | 0.002 |
| Science and technology studies | 0.001 | 0.001 |
| Scholarly communication | 0.002 | 0.001 |
| Open science | 0.002 | 0.001 |
| Research integrity | 0.001 | 0.002 |
| Insufficient payload (model declined to judge) | 0.002 | 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".