Defining an independent reference model for event detection skill scores
Bibliographic record
Abstract
Event detection analysis is a data-model comparison technique in which all observational and numerical values are converted into a binary yes-no categorization of being in or out of ”event state.” Common metrics within event detection analysis include several skill scores – specifically those of Heidke, Peirce (true skill statistic), Clayton, and Gilbert (equitable threat score). All of these skill scores use the general skill score formula, comparing a metric score for the new model against that of a reference model. Moreover, all of them use, to some degree, the same ”expected random matrix” as the reference model. This matrix reshuffles the two number sets of observed and modeled events, randomizing when events occur. These skill scores are, therefore, based on the new model results and thus depend on its performance. That is, these are not calculated relative to an independent reference model. It is shown that for a given metric score (holding proportion correct or critical success index constant), these skill scores have a range of possible values. Conversely, identical skill scores could result from a range of original metric scores. It is recommended to stop using these named skill scores and instead use one of the presented alternatives. One reference model option uses the observed events in place of the new modeled events, while the other uses a 50-50 ”coin flip” option (i.e., truly random chance). These new skill score formulas map one-to-one with the underlying metric values and are, therefore, appropriate for inter-model comparison or intra-model assessment.
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 | 0.004 | 0.002 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.000 |
| Bibliometrics | 0.001 | 0.000 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.000 | 0.001 |
| 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".