Setting BirdNET confidence thresholds: species-specific vs. universal approaches
Bibliographic record
Abstract
BirdNET is widely used in avian acoustic research, providing species predictions alongside confidence values that represent the algorithm’s certainty in species identification. Setting thresholds for these confidence values can increase precision (i.e., the percentage of true positives out of all the identified predictions) but may decrease the predictions retained and exclude true positives that fall below the threshold. This study evaluates two methods for setting confidence thresholds using a two-year audio dataset from western Canada, focusing on 19 target species: (1) a universal threshold of 0.7 across all species and (2) species-specific thresholds defined as the minimum confidence required to achieve a precision of at least 0.9. The universal threshold yielded precision ranging from 0.7 to 1.0 across species but retained only 17 ± 14% (SE) of BirdNET predictions. In contrast, species-specific thresholds ensured precision above 0.9 while retaining 70 ± 37% (SE) of predictions. Species-specific thresholds varied across species but were generally lower than 0.35. Confidence values associated with BirdNET predictions were found to be species specific, but no clear link was observed between BirdNET’s performance and species' song/call complexity, defined as song duration, bandwidth, and number of inflections. Our results confirm that species-specific thresholds offer higher precision and retain more predictions compared to a universal threshold. We provide a step-by-step workflow, including R code, to help researchers define species-specific thresholds that ensure reliable interpretation of BirdNET outputs. Additionally, we discuss how our workflow aligns with and complements previously proposed approaches for setting BirdNET thresholds.
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 | 0.000 | 0.000 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.000 | 0.000 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.001 | 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 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".