摘要
建立高精度水量预测算法模型,有利于水资源充分利用。以北京市2002-2015年需水量为例,对数据进行相关性分析后选出主要影响因素,然后采用主成分回归法、逐步回归法、灰色模型以及BP神经网络共4种方法进行建模,并用北京市2016年和2017年数据进行模型精度验证。结果表明:4种方法都适合用于城市需水量预测,其中主成分分析和逐步回归分析两种方法主要考虑了多元线性回归存在多重共线性,但是逐步回归模型优于主成分回归模型。将4种模型进行对比验证,BP神经网络模型预测精度最高,平均相对误差达到0.79%,用来预测2016-2017年需水量,预测结果分别为38.66亿m^3、39.49亿m^3,适合作为城市需水量预测方法。
Establishing a high-precision water quantity prediction algorithm model is conducive to the full use of water resources.The article takes the water demand data from 2002 to 2015 in Beijing as an example,and analyzes the data to select the main influencing factors.Then,using principal component regression method,stepwise regression method,grey model and BP neural network,the four methods are used for modeling;and the city's 2016 and 2017 data are used to verify the accuracy of the model.According to the results,these four methods are suitable for urban water demand forecasting.Among them,principal component analysis and stepwise regression analysis are mainly used to consider the existence of multiple collinearity in multiple linear regression,but the final stepwise regression model is superior to the principal component regression model.After comparing these models,the BP neural network model has the highest prediction accuracy,and the average relative error is 0.79%.After comparing these models,the BP neural network model has the highest prediction accuracy,and the average relative error is 0.79%.It is used to predict the water demand of the city from 2016 to 2017.The predicted results are 3.866 billion m^3 and 3.949 billion m^3 respectively.It is suitable as a method for predicting urban water demand.
作者
聂红梅
赵建军
李兴菊
王迎
NIE Hong-mei;ZHAO Jian-jun;LI Xing-ju;WANG Ying(College of Science,Kunming University of Science and Technology,Kunming 650500,China)
出处
《软件导刊》
2019年第10期69-73,共5页
Software Guide
基金
国家自然科学基金青年科学基金项目(11103069)
关键词
需水量预测
多重共线性
主成分回归模型
逐步回归模型
灰色模型
BP神经网络模型
water demand forecast
multicollinearity
principal component regression model
stepwise regression model
gray model
BP neural network model
