The Short-term Predicting Method of Algal Blooms Based on Libsvm and Elman Neural Network Modeling

Abstract

After the major reasons of water bloom were analyzed, using the rough set theory and principal component analysis respectively to identify the main factors affecting the forecast algal blooms. On this basis, to take advantage of the Libsvm water bloom prediction model and Elman water bloom prediction model for the short-term prediction of algal blooms phenomenon respectively. Obtained through the fitting networks in the long-term forecasting of algal blooms, the Libsvm prediction accuracy is much higher than the prediction accuracy of artificial neural network. And the Elman neural network can predict the variation of chlorophyll in short-term well, which laid the foundation for in-depth study of the short-term water blooms prediction methods, since the Elman neural network ability of generalization is stronger, of network prediction is more accurate, of fitting performance is better.

Keywords

• Water bloom
• Libsvm
• Elman
• Prediction
• Modeling.