Read the full-text of this article for free

Title: Research on green trade data prediction under global economic shock based on ConvLSTM model oriented towards reducing carbon emissions

Authors: Nianjie Shang; Yue Zhang; Yan Zhang; GaiRong Dai

Addresses: Trade Management Professional, Management, Woosuk University, Jeonju, 55338, North Jeolla Special Autonomous Province, South Korea ' Trade Science Major, Trade Science and Trade Discipline, Silla University, Busan, 46958, Gyeongsangnam-do, South Korea ' Trade Management Professional, Management, Woosuk University, Jeonju, 55338, North Jeolla Special Autonomous Province, South Korea ' Landscape Architecture and Civil Engineering, Landscape Architecture Discipline, Woosuk University, Jeonju, 55338, North Jeolla Special Autonomous Province, South Korea

Abstract: Existing green trade data prediction models focus only on the temporal characteristics of the data, while ignoring the spatial relationships of the data, resulting in large prediction errors for trade volume (M and X). This paper takes Sino-Korean trade as the main research object, and uses the convolutional long short-term memory (ConvLSTM) model to predict trade volume (M and X) data by combining the advantages of spatiotemporal features. This paper first collects and preprocesses relevant green trade data, then constructs a ConvLSTM model, and finally uses the model to output the predicted values of trade volumes M and X for the next year and compares them with the actual data. Experimental results show that the RMSE and MAE of the ConvLSTM model are 16,300 and 20,500, respectively, which are 1900 and 2300 lower than those of the LSTM model.

Keywords: economic shock; green trade; data prediction; ConvLSTM model; spatial features; low-carbon economy; trade volume.

DOI: 10.1504/IJEP.2026.151758

International Journal of Environment and Pollution, 2026 Vol.76 No.1/2, pp.116 - 139

Received: 21 May 2025
Accepted: 31 Oct 2025
Published online: 18 Feb 2026 *