PCA-RBF网络预测模型在水质信息预测中的应用

Application of PCA-RBF Network Model in Prediction of the Quality of Water Information

在分析RBF神经网络预测算法和主成分分析方法的基础上,本文针对地理空间数据的复杂相关性和较强非线性,深入研究了主成分分析与RBF神经网络结合原理,构建了PCA-RBF网络预测模型,最后将预测模型应用于水质信息预测中。实验结果表明:该模型对海量的高维异构数据可进行有效降维,从而优化RBF神经网络结构,有效地提高了地理空间数据预测时的精度,并为GIS领域地理空间数据的预测提供了一种新思路。

Based on the analysis of RBF neural network prediction algorithm and principal component analysis method, this paper aimed at complex correlation and strong nonlinearity of geographical spatial data, in - depth study of the combine principle of principal component analysis and RBF neural network, constructed the PCA - RBF network prediction model. Finally, the model is applied to predict water quality information, the experimental results show that, in the high dimensional heterogeneous data, the model can effec- tively reduce the dimensionality, so as to optimize the structure of RBF neural network, and effectively improves the prediction accura- cy of the geographic spatial data, and provides a new idea for predicting geospatial data in GIS.