Research on One Intelligent Prediction Method for Water Bloom

An intelligent method on short-term prediction on water bloom of BP neural network based on rough set and wavelet analysis is proposed in this paper. This method analyzes factors of effecting the outbreak of water bloom, and these many factors which were processed by reduction method based on rough set were used as input information of the prediction model; after analyzing the main input information by wavelet multi-resolution, it can eliminate the interference factors in the input information, and use BP network to establish the non-linear relationship between input factors and result of water bloom prediction. After experimental simulation, it can validate that this kind of short-term forecast model can predict the short-term change regularity of chlorophyll more precisely, and provides an efficient new method for short-term prediction of water bloom.