生活用水量组合预测模型及其应用

生活用水量是一个复杂的经济系统,采用单个预测模型很难准确地反映其实际情况,结合全国1998～2010年的生活用水量数据,分别采用自回归模型、指数平滑模型、改进的GM(1,1)模型预测全国生活用水量。通过冗余检验,自回归模型、指数平滑模型、改进的GM(1,1)模型均不是冗余方法,并对这三种预测模型采用极小化误差平方和的方法确定最优加权系数,建立了生活用水量的组合预测模型。结果表明,与单项预测模型相比,该组合预测模型具有更高的精度和更好的适应性,且预测结果更加可靠。

Combination Prediction Model of Domestic Water Consumption and Its Application

Domestic water consumption is a complex economic system. It is difficult to accurately reflect actual situation when taking a single forecasting model. Combining with the domestic water consumption in China during 1998-2010, autoregressive model, exponential smoothing model and improved GM(1,1) model are used to predict the domestic water consumption in China, respectively. By making redundancy check, it finds out that the autoregressive model, exponential smoothing model and improved GM(1,1)model are not redundancy methods. Therefore, the minimum sum of squares of forecasting error is applied to determine the optimal weight coefficient and combination prediction model for the domestic water consumption in China is established. The results indicate that the combination prediction model has higher precision and better adaptability with more reliability.