基于P-ISSA-GRU模型的养殖水体溶解氧含量预测

打开文本图片集

中图分类号：TP391 文献标识码：A 文章编号：1000-4440（2025）09-1781-10

Abstract：Inorder to solve the problem of low prediction accuracy of dissolved oxygen（DO）content in aquaculture water，thisstudyproposedagatedrecurrentunit（GRU）predictionmodelbasedonanimproved sparrowsearchalgorithm （ISSA）.Thecorrlationcoefficients ofvarious factors in waterand dissolvedoxygencontent weredeterminedbyPearsoncorrelationcoefficient method，and thestrong corelation factors wereselecedas the inputfeaturesofte model.Byintroducing Tent chaotic mapping to improve population initialization，adaptive dynamic weight factor ω to dynamically change the weight coeficient，Gaussianperturbation（GP）toimprovetheoptimal locationupdating，theabilityof sparrowsearchalgorithm （SSA）to findglobalandlocaloptimalsolutionswasenhanced，and itsconvergencespeedwas increased.Anonlineardissolved oxygen content prediction model （P-ISSA-GRU）was constructed by optimizing the GRU network using ISSA for parametersearch.The experimental resultsshowedthattheP-ISSA-GRU model exhibited superior prediction performance com

pared to five other commonly used models.The mean square error（MSE）wasO.152（mg/L）²，themeanabsoluteerror （ MAE ）was 0.311mg/L ，the root mean square error （RMSE）was 0.390mg/L ，andthe coefficientofdetermination （ R2 ）was O.984.Therefore，compared to traditional models，the P-ISSA-GRUmodel developed in this study demonstratesimproved predictiveperformance for dissolved oxygen contentin aquaculturewater.

Key words：prediction of dissolved oxygen content；Pearson correlation coeffcient； improved sparrow search algorithm（ISSA）；gated recurrent unit （GRU）

溶解氧是衡量池塘水体生态参数变化的重要指标，也是影响养殖池塘环境质量的关键因子[1-2]。（剩余12343字）