Fishermen’S Expert Views on the Causes of Fish Poisoning in Fiji: An Investigation through Data Mining Technique

摘要

Fish poisoning is a common cause of human illness in Fiji. Poisonous substances can be initially ingested by fish that typically feast upon marine green growth, spilt oil, dirty water and silted rubbish. Humans can fall ill when eating these infected fish. This is of particular concern given that fish is the staple diet of most Fijians. This research presents fishermen's views on the significant causes and risk factors of fish poisoning in different regions of Fiji: Rakiraki, Tavua, Sonaisaili Island, Gua Island, Nandi, Lautoka, Suva. The study used a computational intelligence-based (CI), data mining approach to elucidate the fishermen's views. Association rule mining (ARM) was then utilised to find correlations and associations between these views and to determine the perceived main causes of fish poisoning. The aim of this research was to explore the efficacy of an ARM approach (and the development of an appropriate database) in determining the common views of expert fishermen on the causes and risk factors that contribute to fish poisoning. The results indicated that the fishermen held common beliefs of causal relationships between various environmental factors and fish poisoning. These included contaminated passageways of fish, velella areas, water vehicle pollution, water temperatures, dirty and polluted water, heavy rains and flooded waters, the summer season, algal blooms, deep-sea environments and irresponsible human activities. Conceivably, the outcomes of this study may aid designing fish poisoning diagnostic systems (including early diagnosis of poisoning in fish) and through reducing or preventing predicating factors in the first instance.