Integrating Artificial Neural Networks into the Gravity Model: Analyzing Apparel Export from Belt and Road Asian Countries to the United States

Document Type

Journal Article

Publication Date

2025

Keywords

Belt and Road Initiative; Gravity Trade Model; Panel Data Regression; Artificial Neural Network; Extra Trees Regression analysis

DOI

10.33552/JTSFT.2025.11.000766

Abstract

Launched by China in 2013, the Belt and Road Initiative (BRI) aims to enhance trade connectivity between Asia and the global market while fostering mutually beneficial economic cooperation among participant countries. While previous research has highlighted the positive effects of the BRI on trade between China and its partner nations along the Belt and Road (B&R), there is limited understanding of its impact on trade among B&R countries themselves. This study seeks to fill this gap by examining bilateral apparel trade patterns, specifically focusing on clothing exports from China and 21 Asian B&R countries/cities to the United States (US). To analyze these trade flows, we employed a refined gravity trade model augmented with the Logistics Performance Index (LPI) and utilized an Artificial Neural Network (ANN) framework. Additionally, an Extra Trees Regression (ETR) analysis was conducted to explore the factors influencing apparel export values to the US. Data spanning from 2008 to 2024 across the 22 countries/cities were analyzed. The findings indicate that the BRI trade corridor significantly impacts apparel exports from China and other developing B&R Asian countries to the US. Notably, the ANN results indicate that incorporating the LPI feature does not enhance model performance. Furthermore, our results demonstrate that the ANN offers superior predictive accuracy compared to traditional regression models, with ETR serving as a valuable complementary tool to enhance the ANN analysis. Future research can expand on this work by exploring the application of ANN and ETR in new contexts to examine the BRI trade flow from other countries in the Far East and broader regions.

Source Publication

Journal of Textile Science & Fashion Technology

Volume Number

11

Issue Number

4

ISSN

2641-192X

This document is currently not available here.

Link to Publisher’s Version

Share

COinS