Assessment of Machine Learning Models to Identify Port Jackson Shark Behaviours Using Tri-Axial Accelerometers

Abstract

Movement ecology has traditionally focused on the movements of animals over large time scales, but, with advancements in sensor technology, the focus can become increasingly fine scale. Accelerometers are commonly applied to quantify animal behaviours and can elucidate fine-scale (<2 s) behaviours. Machine learning methods are commonly applied to animal accelerometry data; however, they require the trial of multiple methods to find an ideal solution. We used tri-axial accelerometers (10 Hz) to quantify four behaviours in Port Jackson sharks (Heterodontus portusjacksoni): two fine-scale behaviours (<2 s)—(1) vertical swimming and (2) chewing as proxy for foraging, and two broad-scale behaviours (>2 s–mins)—(3) resting and (4) swimming. We used validated data to calculate 66 summary statistics from tri-axial accelerometry and assessed the most important features that allowed for differentiation between the behaviours. One and two second epoch testing sets were created consisting of 10 and 20 samples from each behaviour event, respectively. We developed eight machine learning models to assess their overall accuracy and behaviour-specific accuracy (one classification tree, five ensemble learners and two neural networks). The support vector machine model classified the four behaviours better when using the longer 2 s time epoch (F-measure 89%; macro-averaged F-measure: 90%). Here, we show that this support vector machine (SVM) model can reliably classify both fine- and broad-scale behaviours in Port Jackson sharks.

Internet on animals: Wi‐Fi‐enabled devices provide a solution for big data transmission in biologging

“…It is the first time that a biologging tag implements a modular software architecture (to the best of our knowledge), simplifying the integration of new sensors and onboard processing algorithms based on fine‐scale data. This could include instantaneous orientation estimation based on IMU data (see Supporting Information for details), dead reckoning (Bidder et al., 2015; Gunner et al., 2021), behaviour detection by machine learning (Kadar et al., 2020; Rast et al., 2020) or lossy compression (Nuijten et al., 2020). With the 240 MHz multi‐CPU microcontroller onboard the WildFi, it is possible to run such algorithms in parallel and in real time.…”

Classifying Goliath Grouper (Epinephelus itajara) Behaviors from a Novel, Multi-Sensor Tag

“…In a study investigating the performance of eight conventional machine learning methods classifying acceleration data into behavioral classes for Port Jackson sharks ( Heterodontus portusjacksoni ), the SVM and RF performed best, using 2 s epochs for labeling the data. The two methods obtained equal overall accuracy (89%) but the SVM achieved superior performance for fine-scale behaviors such as chewing [ 7 ]. Conversely, RFs performed better than SVMs for classifying acceleration data obtained from Griffon vultures ( Gyps fulvus ) into seven behaviors [ 6 ].…”

“…Time domain summary statistics included average, standard deviation, minimum, maximum, median, skewness, kurtosis, median absolute deviation, inverse covariance, and interquartile range. Summary statistics were also calculated for overall dynamic body acceleration (ODBA) [ 6 , 7 , 8 , 51 , 52 ]. The accelerometer records total acceleration which comprises the gravitational component of acceleration (which reflects tag orientation, and thus animal posture, in relation to the earth’s gravitational pull) and dynamic acceleration caused by the animals’ body movement.…”

“…Two supervised ML algorithms—a random forest (RF) and a SVM—were built using MATLAB 2019a. Both algorithms have been commonly employed to recognize behavior from acceleration data obtained from numerous species [ 6 , 7 , 56 , 57 , 58 ]. Ensemble classifiers, such as RFs, combine predictions from multiple base estimators to make a more robust model.…”

“…However, with each sensor potentially yielding millions of data points, manually deciphering behaviors from these inertial measurement unit (IMU) data sets is impractical. As such, numerous machine learning (ML) methods have been employed to automate the process of classifying animal-borne sensor output into behavioral classes [6][7][8][9].…”