Applications of machine learning in animal behaviour studies

Valletta, John Joseph; Torney, Colin J.; Kings, Michael; Thornton, Alex; Madden, Joah R.

doi:10.1016/j.anbehav.2016.12.005

Cited by 406 publications

(312 citation statements)

References 70 publications

Supporting

Mentioning

309

Contrasting

Unclassified

Order By: Relevance

“…Because the main objective of the present study is to implement decision trees based on clustering methods, we first categorized the carcasses in accordance with their ribeye steak tenderness scores . Thus, we used different unsupervised learning tools such as k ‐means, hierarchical cluster analysis (HCA) and partitioning around medoids clustering algorithms . The tenderness scores were mean centred prior to any clustering analysis by computing Z ‐scores.…”

Section: Methodsmentioning

confidence: 89%

“… K‐ means is an iterative clustering method used to automatically partition a dataset into k groups where the number of clusters k is assumed to be fixed a priori by minimizing relative distances . This algorithm consists of two separate phases: the first phase is to define k centroids, one for each cluster, and the second phase is to take each point belonging to the given dataset and associate it with the nearest centroid. HCA is a cluster analysis that seeks to build a hierarchy and binary tree of clusters . It classifies a dataset into groups that are internally homogeneous and externally isolated on the basis of measuring the similarity or dissimilarity between groups. …”

Section: Methodsmentioning

confidence: 99%

See 1 more Smart Citation

Decision tree, a learning tool for the prediction of beef tenderness using rearing factors and carcass characteristics

2018

View full text Add to dashboard Cite

show abstract

Section: Methodsmentioning

confidence: 89%

Section: Methodsmentioning

confidence: 99%

Decision tree, a learning tool for the prediction of beef tenderness using rearing factors and carcass characteristics

2018

View full text Add to dashboard Cite

show abstract

“…So far our approach focused on detecting specific morphological traits of the insect and developing a computer vision algorithm able to discern between insects following a “supervised parametric learning” (Crisci, Ghattas, & Perera, ). A different strategy, based on “machine learning algorithms,” will be explored in the future (Valletta, Torney, Kings, Thornton, & Madden, ). Developments on image analysis are expected to enhance the efficacy of the trapping system, further reducing the operational cost and ultimately reaching a completely automatized FF trapping and identification system.…”

Section: Discussionmentioning

confidence: 99%

Electronic traps for detection and population monitoring of adult fruit flies (Diptera: Tephritidae)

Shaked

Amore²,

Ioannou

et al. 2017

J Applied Entomology

View full text Add to dashboard Cite

During the last decades, the economic importance of tephritid fruit flies (FF) has increased worldwide because of recurrent invasions and expansions into new areas, and reduced control capabilities of current control systems. Efficient monitoring systems, thus, are required to provide fast information to act promptly. With this aim in mind, we developed two electronic trap (e‐trap) versions for adult FF: one with specific volatiles for male and female adult Ceratitis capitata, and the second, based on the attraction of adult FF to yellow colour, targeting Dacus ciliatus, Rhagoletis cerasi and Bactrocera oleae. In the case of B. oleae, the female pheromone and ammonium bicarbonate were added as synergists. In the two versions, attracted FF were retained in the trap on glued surfaces. Real‐time images of the surfaces were automatically taken and transmitted to a server. We tested the two e‐trap versions in insect‐proof cages, where flies were released and recaptured, and in commercial orchards throughout the Mediterranean: C. capitata in peach orchards in Italy; R. cerasi in cherry orchards in Greece; B. oleae in olive orchards in Spain and in Greece; and D. ciliatus in melons in plastic tunnels in Israel. The e‐trap showed excellent abilities to transmit real‐time images of trapped FF and a high specificity for trapping different FF species. The ability of the entomologist to correctly classify FF from images in the office was >88%. In addition, average number of flies/trap in e‐trap grids did not differ from numbers reported on grids of conventional traps that were operating simultaneously. The e‐traps developed and tested in this study provide the basis for the real‐time monitoring of FF were no olfactory attractants are available, and for the surveillance of alien FF incursions where generic, but not specific, olfactory attractants exists.

show abstract

“…This more automated form of feature extraction has been successfully applied in speech, audio and image recognition where they have outperformed other machine learning techniques (see LeCun, Bengio, & Hinton, for a review). Deep learning is a relatively new ML technique that to our knowledge has not been applied to animal tracking data or animal behavioural studies to date, but that has been suggested to be a potentially useful tool (Valletta, Torney, Kings, Thornton, & Madden, ).…”

Section: Introductionmentioning

confidence: 99%

Predicting animal behaviour using deep learning: GPS data alone accurately predict diving in seabirds

et al. 2017

View full text Add to dashboard Cite

Abstract1. To prevent further global declines in biodiversity, identifying and understanding key habitats is crucial for successful conservation strategies. For example, globally, seabird populations are under threat and animal movement data can identify key at-sea areas and provide valuable information on the state of marine ecosystems. To date, in order to locate these areas, studies have used global positioning system (GPS) to record position and are sometimes combined with time-depth recorder (TDR) devices to identify diving activity associated with foraging, a crucial aspect of at-sea behaviour. However, the use of additional devices such as TDRs can be expensive, logistically difficult and may adversely affect the animal. Alternatively, behaviours may be resolved from measurements derived from the movement data alone. However, this behavioural analysis frequently lacks validation data for locations predicted as foraging (or other behaviours).2. Here, we address these issues using a combined GPS and TDR dataset from 108 individuals by training deep learning models to predict diving in European shags, common guillemots and razorbills. We validate our predictions using withheld data, producing quantitative assessment of predictive accuracy. The variables used to train these models are those recorded solely by the GPS device: variation in longitude and latitude, altitude and coverage ratio (proportion of possible fixes acquired within a set window of time).3. Different combinations of these variables were used to explore the qualities of different models, with the optimum models for all species predicting non-diving and diving behaviour correctly over 94% and 80% of the time, respectively. We also demonstrate the superior predictive ability of these supervised deep learning models over other commonly used behavioural prediction methods such as hidden Markov models.4. Mapping these predictions provides useful insights into the foraging activity of a range of seabird species, highlighting important at sea locations. These models have the potential to be used to analyse historic GPS datasets and further our understanding of how environmental changes have affected these seabirds over time. K E Y W O R D Scommon guillemot, European shag, foraging, machine learning, prediction, razorbill, time-depth recorder 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.

show abstract

Applications of machine learning in animal behaviour studies

Cited by 406 publications

References 70 publications

Decision tree, a learning tool for the prediction of beef tenderness using rearing factors and carcass characteristics

Decision tree, a learning tool for the prediction of beef tenderness using rearing factors and carcass characteristics

Electronic traps for detection and population monitoring of adult fruit flies (Diptera: Tephritidae)

Predicting animal behaviour using deep learning: GPS data alone accurately predict diving in seabirds

Contact Info

Product

Resources

About