Adaptive Neuro Fuzzy Data Aggregation Model for Developing and Planning for Aquaculture Farming Practices

Abstract

The dynamic nature of environmental elements limits the power and information flow of sensor-based network systems, which are critical to the performance of real-time systems, particularly in the fisheries/aquaculture sector. There is a need for a techniquewhich will improve the network information flow. As a result, a data aggregation process utilising advanced artificial intelligence methods such as Sugeno and mamdanifuzzy system, Adaptive Neuro (or Network based) -Fuzzy Inference System, deep learning, neural networks, and others to reduce the communication activity that creates a single data by aggregating information from a group of various source data in the cluster head. In this backdrop, sugeno fuzzy based adaptive neuro fuzzy data aggregation model/system was developed and validated in aquaculture systems to minimise traffic, improve sensor network efficiency, and create a cost-effective system for the predicted output. It will also be valuable for creating and planning aquaculture farming practises. Results obtained from the models were validated using four statistical parameters. In this model, 70 training dataset and 30 testing dataset were used for validation. The aggregation provides accurate result and has Rootmean square error (0.01936), Coefficient of determination (0.999999245), Mean relative percent error (0.1936)and Variance Account For (99.9837 %). the results of the developed ANFIS model and its tool reveals that will be useful for developing and planning aquaculture farming practices.