基于PCA的模糊BP网络建模方法在藻类生长状态预测中的应用

摘要

In order to describe the complex nonlinear mapping relationship that algae growth is affected by various physical-chemical factors of environment, the fuzzy Back Propagation (BP) network modeling algorithm combined with Principal Component Analysis (PCA) method is adopted to predict the state of algae growth. This method can effectively reduce the dimension of sample data, simplify the complexity of the model system, so that the model has a fast convergence speed and relatively low dimensions. Practical test illustrates that the fuzzy BP network model based on PCA can be well applied to state prediction of algae growth.%为较好地描述藻类生长受环境中各种理化因子作用和影响的复杂的非线性映射关系,将主成分分析(Principal Component Analysis,PCA)与模糊反向传播(Back Propagation,BP)网络模型算法相结合对藻类生长状态进行预测.该方法可有效降低样本数据的维数,简化模型系统的复杂程度,使模型具有较快的收敛速度和相对较低的维数.实际测试表明,基于PCA的模糊BP网络模型能够较好地应用于针对藻类生长的状态预测.