The present invention relates to a fuzzy modeling method for predicting algal blooms, comprising: a first step for receiving input of water quality and hydrological data; a second step for calculating a coefficient of correlation between input/output signals; a third step for applying a wavelet transform analysis technique to filter noise signals from the input/output signals; a fourth step for applying a subtractive clustering algorithm to determine a premise parameter using the input/output signals filtered; a fifth step for optimizing a consequent parameter using a weighted least squares algorithm; a sixth step for returning to the fourth step if an error has a value greater than an allowed error value, and proceeding to the next step if less than the allowed error value; and a seventh step for testing validity of a model.
展开▼
