The SMART technologies that are being developed for on-farm disease diagnosis, and disease and water/milk quality monitoring can be supported with Big Data analytics for discovering new knowledge. In this research, the data that will be generated within the system infrastructure will be utilized using machine learning approaches.
First, after determining baseline disease incidence and milk quality, spatial and temporal deviations from this baseline are determined. Deviations are investigated to define the threshold at which an early warning should be generated.
Second, data from milk samples are used to train machine learning algorithms for the prediction of deviations in milk composition and mastitis.
Third, the same is done for water quality monitoring, where the nutritional status of the pond and activity of the fish is monitored, and early warning systems are installed that identify under- or over-nutrification or deviant fish behavioural patterns.
Fourth, the robustness of different types of ponds (e.g. shrimp versus fish, nutritious ponds versus regular ponds) against temporal and spatial deviations is analysed, making it possible to define the best system under different circumstances.