融合BKA优化的CNN-LSTM模型在地面物流碳排放预测中的应用

打开文本图片集

中图分类号：TP391；TP181 文献标识码：A 文章编号：2096-4706（2025）17-0078-05

Abstract：Inorder tosolvetheproblemsofstrongdata volatilitysignificant nonlinearityandcomplex time dependence inthe predictionofcarbon emissons from ground logistics，this paper proposes a CNN-LSTMdeepprediction model based onBlack-winged KiteOptimizationAlgorithm（BKA）.Themodeluses ConvolutionalNeural Network（CNN） toextractlocal features，employsLong Short-Term Memory （LSTM）tomodel temporal dependencies，andutilizesBKAtoautomatically optimize keyhyperparameters.The experimenttakes thedailycarbonemisiondataofChina's groundlogistics in2024as the researchobject，constructsaslidingwindowtimeseriessample，andcariesoutmodeltrainingandtesting.Theresultsshowthat theproposed modelissuperior totheunoptimized model inMAE，MSE，RMSEandother indicators，andhas highprediction accuracyand stability.This method providesafeasible path forthe inteligent predictionofcarbonemissions from ground logistics，andalsoshowstheapplicationvalueofintellgentoptimizationalgorithminparametertuningofDepLeamingmodel.

（eywords： ground logistics; carbon emission prediction;BKA; CNN-LSTM

0 引言

气候变化已成为全球性重大挑战。（剩余6705字）