城市用水量随机梯度回归分析

Stochastic Gradient Regression Analysis of City Municipal and Domestic Water Consumption

城市生活用水量的预测在城市水资源利用和节约用水规划管理中起着非常重要的作用。采用随机梯度回归方法建立城市生活用水量预测模型,并利用天津市近年来的用水量数据进行验证,同时与支持向量机方法进行比较,结果表明随机梯度回归的预测精度较高,为城市管理部门在采取切实可行的政策和制定科学的水资源规划方面提供可靠依据。

The prediction of city municipal and domestic water consumption plays an important role in utilization of urban water resources. A prediction model of city municipal and domestic water consumption was set up based on stochastic gradient boosting, and applied to water consumption of Tianjin. In comparion with support vector machine, the method of stochastic gradient boosting has higher precision. The study will provide a creditable theoretical support for water consumption programming.