EMPLOYING ARTIFICIAL INTELLIGENCE TECHNIQUES IN DATA ANALYSIS OF PESTICIDE TOXICITY PREDICTION

Increased environmental pollution can be attributed to various factors resulting from different industrial and agricultural activities. In fish, for instance, pollutants (such as pesticides) can be captured from multiple sources, including persistent contaminants of agricultural origin. Consequently, in the last few years, there have been reports of a dramatic increase in the global spreading of pollutants in marine ecosystems. This issue has been linked to climate change and numerous artificial factors. Such pollutants can seriously affect aquatic organisms by generating extremely harmful elements known to cause human disease. Pesticides are highly toxic to fish and other organisms, constituting the food web. The toxicity of a pesticide can cause tissue damage or illness. However, acute toxicity of a pesticide refers to the chemical's capacity to cause damage from a single exposure, typically of short duration. Thus, acute toxicity is determined mainly by visual examination of the dermal tissue followed by biochemical and histological tests. Therefore, this work emphasized the development of a computational toxicology model (Artificial Neural Network, ANN)-based diagnosis support approach as an early warning tool for assessing the toxicological threat of pesticide exposure to cultured fish, complemented with ecological assays and observations.