This project aims to systematically trace the evolution of U.S. trade policies from 1789 to the present using advanced machine learning techniques, particularly natural language processing (NLP). By analyzing legislative bills, we will assess changes in trade policies and their underlying political dynamics over time. A key focus involves developing an NLP-based algorithm to standardize historical tariff items by assigning Harmonized System (HS) codes, converting disparate tariff rates into comparable ad valorem rates, and adjusting past monetary values to current equivalents. The project also envisions creating user-friendly applications, including a mobile tool for real-time HS code classification, facilitating data access and analysis for researchers and policymakers.

I am participating in SuperUROP to gain experience in machine learning research and its applications in historical data analysis. Having previously worked in this lab using NLP algorithms for HS code tariff description matching, I am excited to further my research in this area and also develop tools that offer insights into U.S. trade policies and to gain valuable experience for my future career in research and industry.