基于支持向量机的浮游植物密度预测研究

Research on prediction of phytoplankton's density using support vector machines

利用支持向量机方法对具有非线性及突发性特点的浮游植物密度进行了预测,同时与人工神经网络方法预测的结果进行了比较。结果表明,无论是拟和能力还是预测能力,支持向量机方法都明显优于人工神经网络方法,支持向量机方法比较适合于具有小样本、非线性特点的浮游植物密度预测研究。

The theory of support vector machine was used in the prediction of the phytoplankton＇s density with the characteristics of catastrophic and nonlinearity. Furthermore, the predicted result of support vector machine was compared with the result of artificial neutral network. The results showed that the regressed and predicted result of support vector machine was better than artificial neutral network. The theory of support vector machine was fitted for predicting the phytoplankton＇s density with few data and nonlinear.