In this research, a data agglomeration model for smart aquaculture system has been developed and validated using a sugeno fuzzy based machine learning approach. In order to increase production, machine learning (ML) techniques are employed in fisheries and aquaculture to detect and monitor a variety of environmental parameters, including temperature, salinity, dissolved oxygen, and other factors. In the era of developing technology, these metrics are gathered utilizing sensor-based network systems. The input from various sensors must thus be extracted and combined in the cluster head using sophisticated machine learning techniques such as neural networks, mamdani fuzzy system, sugeno fuzzy system etc. to provide more accurate information. Trapezoidal membership functions are used for defining input variables, such as pollution, organic carbon, salinity, dissolved oxygen and pH. In this, triangular membership function is used for defining the output variable (Aggregation). The input values are aggregated using the logical AND operator, truncation implication, weighted average defuzzification, and 1875 fuzzy rules. This model assigns a grade of poor, good, very good, or excellent (on a scale of 1 to 4), depending on the aggregate area of each aqua site in the datasets. The performance of the sugeno model is validated by aquaculture experts classifying the same datasets. The result shows that 96 % of the outputs of the created fuzzy model accord with the conclusions of the aquaculture expert and also it respond to complete aggregate value range, [1-4]. The results show that the sugeno fuzzy based data aggregation model improves the accuracy and reliability of sensed data for the development of smart aquaculture by the planners and decision makers.

Machine Learning; Sugeno fuzzy inference system; data aggregation model; aquaculture.