Optimising Water Quality Classification in Aquaculture Using a New Parameter Pre-selection Approach

Water stands as a pivotal element in aquaculture, and its quality plays a crucial role in the management of fish farming. The inherent non-linearity, dynamics, and instability of its parameters render it a highly complex system to oversee. Conventional methods for assessing water quality in relation to fish farming often prove inadequate. The integration of technology becomes imperative for successful outcomes. In this context, the utilization of artificial intelligence and machine learning techniques emerges as a promising solution. This paper presents a comparison between several methods of pre-selecting water parameters to optimise the classification of water quality for the culture of marine species. The SFI method selects the parameters Temperature, DO, and pH. The results showed that the SFI-KNN combination is the best. It achieved an accuracy rate equal to 99.87%.