A review of supervised learning methods for classifying animal behavioural states from environmental features
Abstract Accurately predicting behavioural modes of animals in response to environmental features is important for ecology and conservation. Supervised learning (SL) methods are increasingly common in animal movement ecology for classifying behavioural modes. However, few examples exist of applying...
| Main Authors: | , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Wiley
2023-01-01
|
| Series: | Methods in Ecology and Evolution |
| Subjects: | |
| Online Access: | https://doi.org/10.1111/2041-210X.14019 |
| _version_ | 1797766979870785536 |
|---|---|
| author | Silas Bergen Manuela M. Huso Adam E. Duerr Melissa A. Braham Sara Schmuecker Tricia A. Miller Todd E. Katzner |
| author_facet | Silas Bergen Manuela M. Huso Adam E. Duerr Melissa A. Braham Sara Schmuecker Tricia A. Miller Todd E. Katzner |
| author_sort | Silas Bergen |
| collection | DOAJ |
| description | Abstract Accurately predicting behavioural modes of animals in response to environmental features is important for ecology and conservation. Supervised learning (SL) methods are increasingly common in animal movement ecology for classifying behavioural modes. However, few examples exist of applying SL to classify polytomous animal behaviour from environmental features especially in the context of millions of animal observations. We review SL methods (weighted k‐nearest neighbours; neural nets; random forests; and boosted classification trees with XGBoost) for classifying polytomous animal behaviour from environmental predictors. We also describe tuning parameter selection and assessment strategies, approaches for visualizing relationships between predictors and class outputs, and computational considerations. We demonstrate these methods by predicting three categories of risk to bald eagles from colliding with wind turbines using, as predictors, 12 environmental state features associated with 1.7 million GPS telemetry data points from 57 eagles. Of the SL methods we considered, XGBoost yielded the most accurate model with 86.2% classification accuracy and pairwise‐averaged area under the ROC curve of 90.6. Computational time of XGBoost scaled better to large data than any other SL method. We also show how SHAP values integrated in the R package (xgboost) facilitate investigation of variable relationships and importance. For big data applications, XGBoost appears to provide superior classification accuracy and computational efficiency. Our results suggest XGBoost should be considered as an early modelling option in situations where the intent is to classify millions of animal behaviour observations from environmental predictors and to understand relationships between those predictors and movement behaviours. We also offer a tutorial to assist researchers in implementing this method. |
| first_indexed | 2024-03-12T20:33:08Z |
| format | Article |
| id | doaj.art-429ee66087e64e46a244999ea18c995d |
| institution | Directory Open Access Journal |
| issn | 2041-210X |
| language | English |
| last_indexed | 2024-03-12T20:33:08Z |
| publishDate | 2023-01-01 |
| publisher | Wiley |
| record_format | Article |
| series | Methods in Ecology and Evolution |
| spelling | doaj.art-429ee66087e64e46a244999ea18c995d2023-08-01T18:55:41ZengWileyMethods in Ecology and Evolution2041-210X2023-01-0114118920210.1111/2041-210X.14019A review of supervised learning methods for classifying animal behavioural states from environmental featuresSilas Bergen0Manuela M. Huso1Adam E. Duerr2Melissa A. Braham3Sara Schmuecker4Tricia A. Miller5Todd E. Katzner6Department of Mathematics and Statistics Winona State University Winona Minnesota USAU.S. Geological Survey Forest and Rangeland Ecosystem Science Center Corvallis Oregon USABloom Research Inc Los Angeles California USAConservation Science Global, Inc West Cape May New Jersey USAU.S. Fish and Wildlife Service Illinois‐Iowa Field Office Moline Illinois USAWest Virginia University Morgantown West Virginia USAU.S. Geological Survey Forest and Rangeland Ecosystem Science Center Boise Idaho USAAbstract Accurately predicting behavioural modes of animals in response to environmental features is important for ecology and conservation. Supervised learning (SL) methods are increasingly common in animal movement ecology for classifying behavioural modes. However, few examples exist of applying SL to classify polytomous animal behaviour from environmental features especially in the context of millions of animal observations. We review SL methods (weighted k‐nearest neighbours; neural nets; random forests; and boosted classification trees with XGBoost) for classifying polytomous animal behaviour from environmental predictors. We also describe tuning parameter selection and assessment strategies, approaches for visualizing relationships between predictors and class outputs, and computational considerations. We demonstrate these methods by predicting three categories of risk to bald eagles from colliding with wind turbines using, as predictors, 12 environmental state features associated with 1.7 million GPS telemetry data points from 57 eagles. Of the SL methods we considered, XGBoost yielded the most accurate model with 86.2% classification accuracy and pairwise‐averaged area under the ROC curve of 90.6. Computational time of XGBoost scaled better to large data than any other SL method. We also show how SHAP values integrated in the R package (xgboost) facilitate investigation of variable relationships and importance. For big data applications, XGBoost appears to provide superior classification accuracy and computational efficiency. Our results suggest XGBoost should be considered as an early modelling option in situations where the intent is to classify millions of animal behaviour observations from environmental predictors and to understand relationships between those predictors and movement behaviours. We also offer a tutorial to assist researchers in implementing this method.https://doi.org/10.1111/2041-210X.14019behavioural classificationboosted classification treeneural networksrandom forestsupervised learningweighted k‐nearest neighbour |
| spellingShingle | Silas Bergen Manuela M. Huso Adam E. Duerr Melissa A. Braham Sara Schmuecker Tricia A. Miller Todd E. Katzner A review of supervised learning methods for classifying animal behavioural states from environmental features Methods in Ecology and Evolution behavioural classification boosted classification tree neural networks random forest supervised learning weighted k‐nearest neighbour |
| title | A review of supervised learning methods for classifying animal behavioural states from environmental features |
| title_full | A review of supervised learning methods for classifying animal behavioural states from environmental features |
| title_fullStr | A review of supervised learning methods for classifying animal behavioural states from environmental features |
| title_full_unstemmed | A review of supervised learning methods for classifying animal behavioural states from environmental features |
| title_short | A review of supervised learning methods for classifying animal behavioural states from environmental features |
| title_sort | review of supervised learning methods for classifying animal behavioural states from environmental features |
| topic | behavioural classification boosted classification tree neural networks random forest supervised learning weighted k‐nearest neighbour |
| url | https://doi.org/10.1111/2041-210X.14019 |
| work_keys_str_mv | AT silasbergen areviewofsupervisedlearningmethodsforclassifyinganimalbehaviouralstatesfromenvironmentalfeatures AT manuelamhuso areviewofsupervisedlearningmethodsforclassifyinganimalbehaviouralstatesfromenvironmentalfeatures AT adameduerr areviewofsupervisedlearningmethodsforclassifyinganimalbehaviouralstatesfromenvironmentalfeatures AT melissaabraham areviewofsupervisedlearningmethodsforclassifyinganimalbehaviouralstatesfromenvironmentalfeatures AT saraschmuecker areviewofsupervisedlearningmethodsforclassifyinganimalbehaviouralstatesfromenvironmentalfeatures AT triciaamiller areviewofsupervisedlearningmethodsforclassifyinganimalbehaviouralstatesfromenvironmentalfeatures AT toddekatzner areviewofsupervisedlearningmethodsforclassifyinganimalbehaviouralstatesfromenvironmentalfeatures AT silasbergen reviewofsupervisedlearningmethodsforclassifyinganimalbehaviouralstatesfromenvironmentalfeatures AT manuelamhuso reviewofsupervisedlearningmethodsforclassifyinganimalbehaviouralstatesfromenvironmentalfeatures AT adameduerr reviewofsupervisedlearningmethodsforclassifyinganimalbehaviouralstatesfromenvironmentalfeatures AT melissaabraham reviewofsupervisedlearningmethodsforclassifyinganimalbehaviouralstatesfromenvironmentalfeatures AT saraschmuecker reviewofsupervisedlearningmethodsforclassifyinganimalbehaviouralstatesfromenvironmentalfeatures AT triciaamiller reviewofsupervisedlearningmethodsforclassifyinganimalbehaviouralstatesfromenvironmentalfeatures AT toddekatzner reviewofsupervisedlearningmethodsforclassifyinganimalbehaviouralstatesfromenvironmentalfeatures |