【摘要】 在全球贸易数字化转型背景下，浙江省作为中国跨境电商发展的标杆区域，其规模预测对政策制定与产业布局具有重要意义。本研究针对传统预测模型在捕捉跨境电商非线性时序规律中的不足，基于2015-2024年浙江省跨境电商进出口总额、跨境支付市场规模、企业数量及港口吞吐量等多维度数据，构建GRU模型进行十年（2025-2034年）规模预测。结果表明，GRU模型在验证集（2023-2024年）的预测误差（MAPE=4.76%）较低，短期预测精度较高。长期预测显示，浙江省跨境电商进出口总额将从2025年的7226.3亿元增长至2034年的10016.6亿元，增速随规模扩张呈现边际递减趋势。研究将为浙江省跨境电商政策制定与产业升级提供量化依据。更多还原

【Abstract】 Under the background of global trade digital transformation, Zhejiang Province, as a benchmark region for cross-border e-commerce development in China, holds significant importance in its scale prediction for policy formulation and industrial layout. To address the limitations of traditional prediction models in capturing nonlinear temporal patterns of cross-border e-commerce, this study constructs a GRU model based on multi-dimensional data from 2015-2024,including Zhejiang’s cross-border e-commerce import-export volume, total import-export volume, cross-border payment market scale, enterprise quantity, and port throughput, to forecast its development scale over the next decade(2025-2034). The results indicate that the GRU model demonstrates low prediction error(MAPE=4.76%) on the validation set(2023-2024), showing high accuracy in short-term forecasting. Long-term predictions reveal that Zhejiang’s cross-border e-commerce import-export volume will grow from 722.63 billion yuan in 2025 to 1,001.66 billion yuan in 2034, with the growth rate exhibiting a?marginal diminishing trend with scale expansion. The study provides quantitative evidence for policy-making and industrial upgrading of cross-border e-commerce in Zhejiang Province.更多还原