摘要
影响城市日供水量的因素较多,以苏州市吴江区两个水厂2015年的历史数据为基础,建立了以温度、天气和假日为变量的多元线性回归预测模型,并将其用于吴江区2016年3月—2017年4月的水量预测。经过检验,两个水厂预测结果的平均绝对相对误差分别为1.94%和2.78%,其中春节前后预测结果误差偏大,对模型进行优化后预测误差显著降低。
Abstract： The daily water demand of a city is affected by several factors. Based on the historical data in 2015 of the two waterworks in Wujiang District, Suzhou, a model to forecast daily water demand was built. The muhi-factor linear regression model took the variation of temperature, climate, and holi- days as variables. The model was used to predict daily water demand from March 2016 to April 2017. By comparison, the mean absolute relative errors of the two waterworks were 1.94% and 2.78%, respec- tively. The forecast error during the Spring Festival was substantially greater than other times. After opti- mization, the forecast error decreased significantly.
出处
《中国给水排水》
CAS
CSCD
北大核心
2017年第23期141-144,共4页
China Water & Wastewater
关键词
日供水量
预测
多元线性回归模型
daily water demand
forecast
multi-factor linear regression model
