Data Driven Insight Into Fish Behaviour and Their Use for Precision Aquaculture

Aquaculture, or the farmed production of fish and shellfish, has grown rapidly, from supplying just 7% for human consumption in 1974 to more than half 2016. This rapid expansion led growth Precision Aquaculture concept that aims exploit data-driven management production, thereby improving farmer's ability monitor, control, document biological processes farms. Fundamental this paradigm is monitoring environmental animal within a cage, processing those data toward farm insight using models analytics. paper presents an analysis behaviour datasets collected at three salmon farms Norway, Scotland, Canada. Information on were hydroacoustic sensors sampled vertical distribution cage high spatial temporal resolution, while network characterised local site conditions. We present AutoML (or automatic machine learning) tools enables developers with limited science expertise train high-quality specific hand. demonstrate how pipelines can be readily applied aquaculture interrogate quantify primary features explains variance. Results variables such as temperature, wind conditions, hour-of-day important drivers motion all sites. Further, there distinct differences factors influenced in-cage variations driven by water depth ambient conditions (particularly dissolved oxygen). The framework offers transferable approach systems, between