Analyzing U.S. Maritime Trade and COVID-19 Impact Using Machine Learning

Peter R. Abraldes; James Rotella; Partha Mukherjee; Youakim Badr

doi:10.4018/978-1-7998-9220-5.ch021

Analyzing U.S. Maritime Trade and COVID-19 Impact Using Machine Learning

Peter R. Abraldes, James Rotella, Partha Mukherjee, Youakim Badr

Engineering Division (Great Valley)

Research output: Chapter in Book/Report/Conference proceeding › Chapter

Abstract

International trade is crucial to the U.S. economy. When measured by metric tonnage, 75% of the volume is transported by sea. This work explores the variables that impact trade volume, both on the individual and aggregate shipment level. Three supervised machine-learning techniques use data about U.S. maritime trade, U.S. unemployment, international tariff rates, and country exchange rates for the period of 2010 through 2019. The same variables are then used in combination with new information specific to COVID-19 for the period of 2019 and 2020. The tonnage of individual shipments remains unchanged before and after the prevalence of COVID-19. However, the frequency of shipments decreases during 2020, resulting in an aggregate decline of trade. Metrics that measure COVID-19's impact produce better model prediction accuracy.

Original language	English (US)
Title of host publication	Encyclopedia of Data Science and Machine Learning
Publisher	IGI Global
Pages	323-341
Number of pages	19
ISBN (Electronic)	9781799892212
ISBN (Print)	9781799892205
DOIs	https://doi.org/10.4018/978-1-7998-9220-5.ch021
State	Published - Jan 1 2022

All Science Journal Classification (ASJC) codes

General Computer Science

UN SDGs

This output contributes to the following UN Sustainable Development Goals (SDGs)

Access to Document

10.4018/978-1-7998-9220-5.ch021

Cite this

@inbook{f5fb0ed60681430bb016d9d72276cbf9,

title = "Analyzing U.S. Maritime Trade and COVID-19 Impact Using Machine Learning",

abstract = "International trade is crucial to the U.S. economy. When measured by metric tonnage, 75% of the volume is transported by sea. This work explores the variables that impact trade volume, both on the individual and aggregate shipment level. Three supervised machine-learning techniques use data about U.S. maritime trade, U.S. unemployment, international tariff rates, and country exchange rates for the period of 2010 through 2019. The same variables are then used in combination with new information specific to COVID-19 for the period of 2019 and 2020. The tonnage of individual shipments remains unchanged before and after the prevalence of COVID-19. However, the frequency of shipments decreases during 2020, resulting in an aggregate decline of trade. Metrics that measure COVID-19's impact produce better model prediction accuracy.",

author = "Abraldes, {Peter R.} and James Rotella and Partha Mukherjee and Youakim Badr",

year = "2022",

month = jan,

day = "1",

doi = "10.4018/978-1-7998-9220-5.ch021",

language = "English (US)",

isbn = "9781799892205",

pages = "323--341",

booktitle = "Encyclopedia of Data Science and Machine Learning",

publisher = "IGI Global",

}

TY - CHAP

T1 - Analyzing U.S. Maritime Trade and COVID-19 Impact Using Machine Learning

AU - Abraldes, Peter R.

AU - Rotella, James

AU - Mukherjee, Partha

AU - Badr, Youakim

PY - 2022/1/1

Y1 - 2022/1/1

N2 - International trade is crucial to the U.S. economy. When measured by metric tonnage, 75% of the volume is transported by sea. This work explores the variables that impact trade volume, both on the individual and aggregate shipment level. Three supervised machine-learning techniques use data about U.S. maritime trade, U.S. unemployment, international tariff rates, and country exchange rates for the period of 2010 through 2019. The same variables are then used in combination with new information specific to COVID-19 for the period of 2019 and 2020. The tonnage of individual shipments remains unchanged before and after the prevalence of COVID-19. However, the frequency of shipments decreases during 2020, resulting in an aggregate decline of trade. Metrics that measure COVID-19's impact produce better model prediction accuracy.

AB - International trade is crucial to the U.S. economy. When measured by metric tonnage, 75% of the volume is transported by sea. This work explores the variables that impact trade volume, both on the individual and aggregate shipment level. Three supervised machine-learning techniques use data about U.S. maritime trade, U.S. unemployment, international tariff rates, and country exchange rates for the period of 2010 through 2019. The same variables are then used in combination with new information specific to COVID-19 for the period of 2019 and 2020. The tonnage of individual shipments remains unchanged before and after the prevalence of COVID-19. However, the frequency of shipments decreases during 2020, resulting in an aggregate decline of trade. Metrics that measure COVID-19's impact produce better model prediction accuracy.

UR - http://www.scopus.com/inward/record.url?scp=85188137603&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85188137603&partnerID=8YFLogxK

U2 - 10.4018/978-1-7998-9220-5.ch021

DO - 10.4018/978-1-7998-9220-5.ch021

M3 - Chapter

AN - SCOPUS:85188137603

SN - 9781799892205

SP - 323

EP - 341

BT - Encyclopedia of Data Science and Machine Learning

PB - IGI Global

ER -

Analyzing U.S. Maritime Trade and COVID-19 Impact Using Machine Learning

Abstract

All Science Journal Classification (ASJC) codes

UN SDGs

Access to Document

Other files and links

Fingerprint

Cite this