RBF神经网络在河流营养盐预测中的应用

Application of RBF neural network in river nutritive salt prediction

根据已监测到的水质数据进行预测一直是河流水质管理的重要组成部分,其中河流营养盐浓度是影响水质的根本因素。文中研究RBF神经网络在河流营养盐浓度预测中的适用性,并与传统的时间序列预测模型:ARIMA进行比较。以朱衣河为研究对象,对河流营养盐主要成分之一的磷酸盐浓度进行预测。通过采集到的时间序列数据对两种预测模型进行仿真,并通过平均误差和均方误差的比较,证明基于RBF神经网络的时间序列预测模型具有较强的预测精度和良好的推广价值能力,在河流营养盐预测中有较高的实用性。

River nutritive salt concentration is a fundamental factor affecting water quality.The applicability of RBF neural network in river nutrient concentration prediction is studied and compared with traditional time series prediction model:ARIMA.Taking Zhuyi River as the research object,the phosphate concentration is predicted,which is one of the main components of riv?er nutritive salt.The simulation test of the two prediction models was carried out with the collected data of time series.By com?paring average error with mean square error,it is proved that the time series prediction model based on RBF neural network has high prediction accuracy,high romotional value and better applicability in prediction of river nutritive salt.