Bio-logger Ethogram Benchmark: A benchmark for computational analysis of animal behavior, using animal-borne tags

  1. Hoffman, Benjamin 1
  2. Cusimano, Maddie 1
  3. Baglione, Vittorio 2
  4. Canestrari, Daniela 2
  5. Chevallier, Damien 3
  6. DeSantis, Dominic L. 4
  7. Jeantet, Lorène 5
  8. Ladds, Monique A. 6
  9. Maekawa, Takuya 7
  10. Mata-Silva, Vicente 8
  11. Moreno-González, Víctor 2
  12. Trapote, Eva 2
  13. Vainio, Outi 9
  14. Vehkaoja, Antti 10
  15. Yoda, Ken 11
  16. Zacarian, Katherine 1
  17. Friedlaender, Ari 12
  1. 1 Earth Species Project
  2. 2 Universidad de León
    info

    Universidad de León

    León, España

    ROR https://ror.org/02tzt0b78

  3. 3 Centre national de la recherche scientifique Borea
  4. 4 Georgia College & State University
    info

    Georgia College & State University

    Milledgeville, Estados Unidos

    ROR https://ror.org/05kr7zx86

  5. 5 African Institute for Mathematical Sciences, Stellenbosch University
  6. 6 Department of Conservation, New Zealand
  7. 7 Osaka University
    info

    Osaka University

    Osaka, Japón

    ROR https://ror.org/035t8zc32

  8. 8 University of Texas, El Paso
  9. 9 University of Helsinki
    info

    University of Helsinki

    Helsinki, Finlandia

    ROR https://ror.org/040af2s02

  10. 10 Tampere University
  11. 11 Nagoya University
    info

    Nagoya University

    Nagoya, Japón

    ROR https://ror.org/04chrp450

  12. 12 University of California, Santa Cruz
    info

    University of California, Santa Cruz

    Santa Cruz, Estados Unidos

    ROR https://ror.org/03s65by71

Editor: Zenodo

Año de publicación: 2024

Tipo: Dataset

Resumen

This repository contains the datasets and experiment results presented in our arxiv paper: B. Hoffman, M. Cusimano, V. Baglione, D. Canestrari, D. Chevallier, D. DeSantis, L. Jeantet, M. Ladds, T. Maekawa, V. Mata-Silva, V. Moreno-González, A. Pagano, E. Trapote, O. Vainio, A. Vehkaoja, K. Yoda, K. Zacarian, A. Friedlaender, "A benchmark for computational analysis of animal behavior, using animal-borne tags," 2023. Standardized code to implement, train, and evaluate models can be found at https://github.com/earthspecies/BEBE/.  Please note the licenses in each dataset folder. Zip folders beginning with "formatted": These are the datasets we used to run the experiments reported in the benchmark paper.  Zip folders beginning with "raw": These are the unprocessed datasets used in BEBE. Code to process these raw datasets into the formatted ones used by BEBE can be found at https://github.com/earthspecies/BEBE-datasets/. Zip folders beginning with "experiments": Results of the cross-validation experiments reported in the paper, as well as hyperparameter optimization. Confusion matrices for all experiments can also be found here. Note that dt, rf, and svm refer to the feature set from Nathan et al., 2012. Results used in Fig. 4 of arxiv paper (deep neural networks vs. classical models){dataset}_ harnet_nogyr{dataset}_CRNN{dataset}_CNN{dataset}_dt{dataset}_rf{dataset}_svm{dataset}_wavelet_dt{dataset}_wavelet_rf{dataset}_wavelet_svm Results used in Fig. 5D of arxiv paper (full data setting)If dataset contains gyroscope (HAR, jeantet_turtles, vehkaoja_dogs):{dataset}_harnet_nogyr{dataset}_harnet_random_nogyr{dataset}_harnet_unfrozen_nogyr{dataset}_RNN_nogyr{dataset}_CRNN_nogyr{dataset}_rf_nogyrOtherwise:{dataset}_harnet_nogyr{dataset}_harnet_unfrozen_nogyr{dataset}_harnet_random_nogyr{dataset}_RNN_nogyr{dataset}_CRNN{dataset}_rf Results used in Fig. 5E of arxiv paper (reduced data setting)If dataset contains gyroscope (HAR, jeantet_turtles, vehkaoja_dogs):{dataset}_harnet_low_data_nogyr{dataset}_harnet_random_low_data_nogyr{dataset}_harnet_unfrozen_low_data_nogyr{dataset}_RNN_low_data_nogyr{dataset}_wavelet_RNN_low_data_nogyr{dataset}_CRNN_low_data_nogyr{dataset}_rf_low_data_nogyr Otherwise:{dataset}_harnet_low_data_nogyr{dataset}_harnet_random_low_data_nogyr{dataset}_harnet_unfrozen_low_data_nogyr{dataset}_RNN_low_data_nogyr{dataset}_wavelet_RNN_low_data_nogyr{dataset}_CRNN_low_data{dataset}_rf_low_data CSV files: we also include summaries of the experimental results in experiments_summary.csv, experiments_by_fold_individual.csv, experiments_by_fold_behavior.csv.  experiments_summary.csv - results averaged over individuals and behavior classesdataset (str): name of datasetexperiment (str): name of model with experiment setting fig4 (bool): True if dataset+experiment was used in figure 4 of arxiv paperfig5d (bool): True if dataset+experiment was used in figure 5d of arxiv paperfig5e (bool): True if dataset+experiment was used in figure 5e of arxiv paperf1_mean (float): mean of macro-averaged F1 score, averaged over individuals in test foldsf1_std (float): standard deviation of macro-averaged F1 score, computed over individuals in test foldsprec_mean, prec_std (float): analogous for precisionrec_mean, rec_std (float): analogous for recallexperiments_by_fold_individual.csv - results per individual in the test foldsdataset (str): name of datasetexperiment (str): name of model with experiment setting fig4 (bool): True if dataset+experiment was used in figure 4 of arxiv paperfig5d (bool): True if dataset+experiment was used in figure 5d of arxiv paperfig5e (bool): True if dataset+experiment was used in figure 5e of arxiv paperfold (int): test fold indexindividual (int): individuals are numbered zero-indexed, starting from fold 1f1 (float): macro-averaged f1 score for this individualprecision (float): macro-averaged precision for this individualrecall (float): macro-averaged recall for this individual experiments_by_fold_behavior.csv - results per behavior class, for each test folddataset (str): name of datasetexperiment (str): name of model with experiment setting fig4 (bool): True if dataset+experiment was used in figure 4 of arxiv paperfig5d (bool): True if dataset+experiment was used in figure 5d of arxiv paperfig5e (bool): True if dataset+experiment was used in figure 5e of arxiv paperfold (int): test fold indexbehavior_class (str): name of behavior classf1 (float): f1 score for this behavior, averaged over individuals in the test foldprecision (float): precision for this behavior, averaged over individuals in the test foldrecall (float): recall for this behavior, averaged over individuals in the test foldtrain_ground_truth_label_counts (int): number of timepoints labeled with this behavior class, in the training set