论文部分内容阅读
根据城市用水量系统具有非线性和随机波动性的特点,为了充分发挥组合灰色神经网络预测模型能够综合单变量预测及非线性处理的优势,同时降低组合权系数计算方法的不确定性对模型预测效果的影响,论文提出了基于马尔科夫链修正的组合灰色神经网络预测模型。将其应用于1980—2009年青海省城市用水量序列的拟合分析,并预测其2010、2015以及2020年的城市需水量。结果表明:基于马尔科夫链修正的组合灰色神经网络预测模型预测结果的误差更小,精度更高。
According to the nonlinear and stochastic volatility of urban water consumption system, in order to give full play to the combined gray neural network prediction model, the advantages of univariate prediction and non-linear processing can be integrated, and the uncertainty of the combination weight coefficient calculation method is reduced. Effect, the paper proposes a combined gray neural network prediction model based on Markov chain correction. Applying it to the fitting analysis of urban water consumption in Qinghai Province from 1980 to 2009 and forecasting its urban water demand in 2010, 2015 and 2020. The results show that the prediction error of the combined gray neural network model based on Markov chain correction is smaller and the accuracy is higher.