Risk Assessment of Whale Entanglement and Vessel Strike Injuries From Case Narratives and Classification Trees
Bibliographic record
Abstract
Entanglements and vessel strikes impact large whales worldwide. Post-event health status is often unknown because whales are seen once or over short spans that conceal long-term health declines. Well-studied populations with high site fidelity verified by photo-ID offer opportunity to confirm deaths, health declines and recoveries. We used known outcome entanglements and vessel strikes of right whales ( Eubalaena glacialis ) and humpback whales ( Megaptera novaeangliae ) to model probabilities of deaths, health declines and recoveries with Random Forest (RF) classification trees. Variables included presence or absence of phrases from case narratives (‘deep laceration’, ‘cyamid’, ‘healing’, ‘superficial’) and a categorical variable for vessel size. Health status post-entanglement was correctly classified in 95.7% of right whale and 93.6% of humpback whale cases (expected by chance=50%). Health status post-vessel strike was correctly classified in 91.4% of right whale and 88.6% of humpback whale cases. Important variables included cyamid presence, emaciation, discolored skin, constricting entanglements, gear-free resightings, superficial or healing lacerations, and vessel size. Cross-validated RF models were applied to unknown outcome cases to estimate the probability of deaths, health declines and recoveries. Total serious injuries (probability of death or health decline > 0.50) assigned by RF were nearly equal to current injury assessment methods applied by biologists for known outcomes. However, RF consistently predicted higher serious injury totals for unknown outcomes, suggesting that current assessment methods may underestimate risk for cases lacking details or long-term observations. Advantages of the RF method include: 1) risk models are based on known outcomes; 2) unknown outcomes are assigned post-event health status probabilities; and 3) identification of important predictor variables improves data collection standards.
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 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 | 0.001 |
| Science and technology studies | 0.000 | 0.001 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.000 | 0.002 |
| 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 itClassification
machine, unvalidatedMachine predicted; a candidate call from one teacher head, 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".