云南水力发电2025，Vol.41Issue(9)：18-23,6.DOI:10.3969/j.issn.1006-3951.2025.09.005
基于数据分解与智能算法优化的岭回归月径流预测
Ridge Regression Monthly Runoff Time Series Prediction Based on Data Decomposition and Intelligent Algorithm Optimization
摘要
Abstract
To improve the accuracy of monthly runoff prediction and solve the problem of collinearity among variables,a monthly runoff time series prediction model is proposed based on wavelet packet decomposition(WPT),which combines escape bird search(EBS)algorithm,pelican optimization algorithm(POA),and ridge regression(RR).The model is tested through monthly runoff prediction examples from the Yingluoxia hydrological station in the Heihe River Basin from 1944 to 2016 and literature review of monthly runoff prediction examples in Laihe River.Firstly,WPT is used to decompose the monthly runoff time series into more regular low-frequency and high-frequency components;Next,briefly introduce the principle of EBS/POA,use EBS/POA to optimize RR ridge parameters,establish WPT-EBS-RR and WPT-POA-RR models,and construct WPT-EBS-RBF and WPT-POA-RBF models based on radial basis function neural network(RBF)for comparative analysis with the unoptimized WPT-RR model;Finally,the predictive performance of the WPT-EBS-RR and WPT-POA-RR models was further validated through a case study of monthly runoff prediction in the Laihe River.The results showed that:The average absolute percentage error of the WPT-EBS-RR and WPT-POA-RR models for predicting monthly runoff from January to December at the Yingluoxia hydrological station ranged from 0.815%to 1.906%and 0.823%to 2.006%,with determination coefficients greater than 0.99,which was superior to other comparative models and had higher prediction accuracy.Using EBS/POA to determine the optimal ridge parameters for RR overcomes the difficulty of determining traditional RR ridge parameters and solves the problem of collinearity among variables.关键词
月径流预测/岭回归算法/逃逸鸟搜索算法/鹈鹕优化算法/小波包变换/岭参数优化Key words
monthly runoff forecast/ridge regression algorithm/escaping bird searchs algorithm/pelican optimization algorithm/wavelet packet transform/ridge parameter optimization分类
建筑与水利引用本文复制引用
潘秀昌,潘思成,崔东文..基于数据分解与智能算法优化的岭回归月径流预测[J].云南水力发电,2025,41(9):18-23,6.基金项目
国家自然科学基金项目(41702278) （41702278）
中国地质调查局地质调查项目(DD20221758,DD20190326) （DD20221758,DD20190326）
