Classifying behavior from short‐interval biologging data: An example with GPS tracking of birds

Bergen, Silas; Huso, Manuela M.; Duerr, Adam E.; Braham, Melissa A.; Katzner, Todd E.; Schmuecker, Sara; Miller, Tricia A.

doi:10.1002/ece3.8395

Cited by 4 publications

(8 citation statements)

References 55 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Likewise, although unsupervised classification is a poor choice for the identification of a priori‐defined behaviors, it can effectively identify patterns or clusters in accelerometer signals, and these can be interpreted as “behavioral modes.” As such, it is reasonable to use post hoc data interpretation to characterize predominant behaviors within those modes. For example, K‐means clustering has been used to identify behavioral “groups” from accelerometer data collected from European shag ( Phalacrocorax aristotelis ) in Scotland, UK (Sakamoto et al, 2009 ), and “movement states” from short‐interval GPS data collected from bald eagles in the midwestern USA (Bergen et al, 2022 ). In situations such as these, use of unsupervised classification is appropriate, since inference is not to specific behaviors, but to “modes” or “states” in which accelerometer data cluster together.…”

Section: Discussionmentioning

confidence: 99%

“…and "movement states" from short-interval GPS data collected from bald eagles in the midwestern USA (Bergen et al, 2022). In situations such as these, use of unsupervised classification is appropriate, since inference is not to specific behaviors, but to "modes" or "states" in which accelerometer data cluster together.…”

Section: Inference From Unsupervised Classification Of Bio-logging Datamentioning

confidence: 99%

See 1 more Smart Citation

Supervised versus unsupervised approaches to classification of accelerometry data

Sur¹,

Hall

Brandt

et al. 2023

Ecology and Evolution

Self Cite

View full text Add to dashboard Cite

Sophisticated animal-borne sensor systems are increasingly providing novel insight into how animals behave and move. Despite their widespread use in ecology, the diversity and expanding quality and quantity of data they produce have created a need for robust analytical methods for biological interpretation. Machine learning tools are often used to meet this need. However, their relative effectiveness is not well known and, in the case of unsupervised tools, given that they do not use validation data, their accuracy can be difficult to assess. We evaluated the effectiveness of supervised (n = 6), semi-supervised (n = 1), and unsupervised (n = 2) approaches to analyzing accelerometry data collected from critically endangered California condors (Gymnogyps californianus). Unsupervised K-means and EM (expectation-maximization) clustering approaches performed poorly, with adequate classification accuracies of <0.8 but very low values for kappa statistics (range: −0.02 to 0.06). The semi-supervised nearest mean classifier was moderately effective at classification, with an overall classification accuracy of 0.61 but effective classification only of two of the four behavioral classes. Supervised random forest (RF) and k-nearest neighbor (kNN) machine learning models were most effective at classification across all behavior types, with overall accuracies >0.81. Kappa statistics were also highest for RF and kNN, in most cases substantially greater than for other modeling approaches. Unsupervised modeling, which is commonly used for the classification of a priori-defined behaviors in telemetry data, can provide useful information but likely is instead better suited to post hoc definition of generalized behavioral states. This work also shows the potential for substantial variation in classification accuracy among different machine learning approaches and among different metrics of accuracy. As such, when analyzing biotelemetry data, best practices appear to call for the evaluation of several machine learning techniques and several measures of accuracy for each dataset under consideration.

show abstract

Section: Discussionmentioning

confidence: 99%

Section: Inference From Unsupervised Classification Of Bio-logging Datamentioning

confidence: 99%

Supervised versus unsupervised approaches to classification of accelerometry data

Sur¹,

Hall

Brandt

et al. 2023

Ecology and Evolution

Self Cite

View full text Add to dashboard Cite

show abstract

“…Some of the highest turbine‐caused mortality rates of this species have been reported in the Midwestern United States, where eagles use both riparian habitats and the upland areas where wind turbines are often built (Schmuecker et al, 2020). Data from GPS telemetry devices on eagles can be used to describe eagle flight behaviour (Bergen et al, 2022), and certain flight behaviours are likely to be associated with higher collision risk. If risky behaviours can be associated with underlying landscape features, there is the potential to identify landscapes where collision risk is relatively more or less likely, thus providing important information to guide turbine siting.…”

Section: Case Studymentioning

confidence: 99%

“…Some of the highest turbine-caused mortality rates of this species have been reported in the Midwestern United States, where eagles use both riparian habitats and the upland areas where wind turbines are often built (Schmuecker et al, 2020). Data from GPS telemetry devices on eagles can be used to describe eagle flight behaviour (Bergen et al, 2022), and certain flight behaviours are likely to be associ-…”

Section: Eagle Flight Datamentioning

confidence: 99%

A review of supervised learning methods for classifying animal behavioural states from environmental features

et al. 2022

Self Cite

View full text Add to dashboard Cite

1. 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. 2. 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.3. 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. 4. 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.

show abstract

“…Moreover, track symmetry and tortuosity can be used to quantitatively assess risk-prone behavior. Previous studies have often described tortuosity as an indicator for collision risk, indicating birds with a more tortuous flight path to be at higher risk for collision [6,38,39]. The inference of these studies is based on the expectation that a more tortuous flight path is associated with a larger amount of time flying; during flight, there will always be a collision risk.…”

Section: Flight Behavior As a Predictor Of Collision Riskmentioning

confidence: 99%

Quantifying Raptors’ Flight Behavior to Assess Collision Risk and Avoidance Behavior to Wind Turbines

et al. 2022

View full text Add to dashboard Cite

Some wind farms have implemented automated camera-based monitoring systems, e.g., IdentiFlight to mitigate the impact of wind turbines on protected birds. These systems have promoted the collection of large amounts of unique data that can be used to describe flight behavior in a novel way. The aim of this study was to evaluate how this unique data can be used to create a robust quantitative behavioral analysis, that can be used to identify risk-prone flight behavior and avoidance behavior and thereby used to assess collision risk in the future. This was achieved through a case study at a wind farm on the Swedish island Gotland, where golden eagles (Aquila chrysaetos), white-tailed eagles (Haliaeetus albicilla), and red kites (Milvus milvus), were chosen as the bird species. These three species are generally rare breeds in Europe and have also been shown to be particularly vulnerable to collisions with wind turbines. The results demonstrate that data from the IdentiFlight system can be used to identify risk-prone flight behaviors, e.g., tortuous flight and foraging behavior. Moreover, it was found that these flight behaviors were affected by both weather conditions, but also their distance to the nearest wind turbine. This data can, thus, be used to evaluate collision risk and avoidance behavior. This study presents a promising framework for future research, demonstrating how data from camera-based monitoring systems can be utilized to quantitatively describe risk-prone behavior and thereby assess collision risk and avoidance behavior.

show abstract

Classifying behavior from short‐interval biologging data: An example with GPS tracking of birds

Abstract: This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited.

Cited by 4 publications

References 55 publications

Supervised versus unsupervised approaches to classification of accelerometry data

Supervised versus unsupervised approaches to classification of accelerometry data

A review of supervised learning methods for classifying animal behavioural states from environmental features

Quantifying Raptors’ Flight Behavior to Assess Collision Risk and Avoidance Behavior to Wind Turbines

Contact Info

Product

Resources

About