Red tides prediction using fuzzy inference and decision tree

A red tide is a temporary natural phenomenon in which harmful algal blooms (HABs) can lead to finfish and shellfish dying en masse. Prediction of red tide bloom consists of a categorical type and a numerical type, which can minimize the mitigation cost of HAB disasters and the suffering caused by the damage from red tide events. The categorical prediction has high precision but it represents a simple binary result, and the numerical prediction can predict how much harm an algal increase causes, but its prediction has lower accuracy than the results of the categorical type. This paper proposes a red tide prediction method that combines fuzzy inference with decision tree to obtain prediction results of the categorical and numerical types. The experimental results demonstrate that the proposed method achieves a better red tide prediction performance than other prediction methods by classifiers.