基于VMD-XGBoost溶解氧预测模型研究

打开文本图片集

中图分类号：S96；TP391 文献标识码：A 文章编号：2095-5553（2025）12-0114-07

Abstract：Adisolved oxygencombination prediction model was establishedfor predicting changes indisolved oxygen in water bodies，providing decision-makingguidance for fieldssuch asaquaculture.The ISSAalgorithm is used to improve theVMDalgorithmand theTSOalgorithm toimprove the XGBoost algorithm，and thedisolvedoxygencombination prediction model combining ISSA—VMD，TSO—XGBoost and ARMA is constructed. The experimental results show thatcompared with the XGBoost model without VMD decomposition，theRMSEis reduced by O.310 2，the R2 is increased by O.084 8，and the ABMAX isreduced by 1.Ol2 3. Furthermore，compared to the EMD—XGBoost， EEMD—XGBoost，CEEMDAN—LSTM，and VMD—LSTM prediction models，the VMD—XGBoost model reduces RMSEby O.160 7，O.0292，O.2796，andO.2747，respectively，and increases R2 by 0.070 4，0.008 2，0.0277，and 0.0151，respectively，indicating that this model offers better prediction accuracy and stability.

keywords：dissolved oxygen；VMD；XGBoost;combined prediction；aquaculture

0 引言

水体溶解氧是评价水体富氧状况的重要指标之一，在水质监测和生态环境管理中具有关键意义[1]。（剩余10353字）