Abstract
SummaryBat echolocation call identification methods are important in developing efficient cost-effective methods for large-scale bioacoustic surveys for global biodiversity monitoring and conservation planning. Such methods need to provide interpretable probabilistic predictions of species since they will be applied across many different taxa in a diverse set of applications and environments. We develop such a method using a multinomial probit likelihood with independent Gaussian process priors and study its feasibility on a data set from an on-going study of 21 species, five families and 1800 bat echolocation calls collected from Mexico, a hotspot of bat biodiversity. We propose an efficient approximate inference scheme based on the expectation propagation algorithm and observe that the overall methodology significantly improves on currently adopted approaches to bat call classification by providing an approach which can be easily generalized across different species and call types and is fully probabilistic. Implementation of this method has the potential to provide robust species identification tools for biodiversity acoustic bat monitoring programmes across a range of taxa and spatial scales.
Highlights
In the face of severe declines in populations of many wildlife species (Butchart et al, 2010; Tittensor et al, 2014) monitoring changes in ecological communities through time and space is critically important for conservation planning and decision making (Magurran et al, 2010)
The most studied methods use call parameters extracted from spectrograms and discriminant function analysis, support vector machines (SVMs) or artificial neural networks are employed for supervised classification (Walters et al, 2012; Parsons and Jones, 2000; Fenton and Bell, 1981)
Andrieu and Roberts (2009) showed that we can obtain samples from the exact posterior by using an estimate of the marginal provided that it is unbiased. This has been exploited by Filippone and Girolami (2014) for binary probit classification with Gaussian processes (GPs) and has been shown to be the most efficient method compared with other Gibbs sampling algorithms
Summary
The most studied methods use call parameters extracted from spectrograms and discriminant function analysis, support vector machines (SVMs) or artificial neural networks are employed for supervised classification (Walters et al, 2012; Parsons and Jones, 2000; Fenton and Bell, 1981) These existing classification tools typically cover a small set of species and point estimates for the model parameters are obtained by using high quality recordings from well-curated collections of bat calls.
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callMore From: Journal of the Royal Statistical Society Series C: Applied Statistics
Disclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
