An Interval Prediction Model for Algal Blooms with Mixed Uncertainty

The uncertainty of bloom outbreak is an important problem in the process of bloom prediction. In this paper, an interregional prediction model for algal blooms with mixed uncertainty is established by considering the uncertainties in the prediction process. Firstly, the Pearson correlation coefficient method was used to extract the main influencing factors from the original data. Then Mallat is used for noise reduction of input variables. The point prediction model combined with quantile regression random forest is used for interval prediction, which effectively avoids the interference of prior assumptions on the interval prediction results. Finally, the model is evaluated by using interval coverage and interval bandwidth. The accuracy and effectiveness of the model are verified by the comparative analysis of the Tamjin River in Korea.