Analysis of the relationship between traditional markets and commodities trade
Date Issued
2023-07
Author(s)
Todorovska, Ana
Milev, Stefan
Vodenska, Irena
T. Chitkushev, Lubomir
Abstract
In a global world, no country, market, or economy
is isolated and can function independently. Interconnectivity
is a fundamental feature of economic systems, including both
traditional financial markets and novel markets. This study aims
to explore the relationships between traditional markets and
commodities trade. We develop a methodology for analyzing the
relationships between five stock market indexes (S&P500, Dow
Jones, BSE, Hang Seng, FTSE) and three commodities (crude
oil, natural gas, gold), based on multimodal publicly available
datasets incorporating structured numerical and unstructured
news and social network data. To find the existence of direc tional associations, we develop an Explainable ML model that
first learns the dependencies between different assets and then
explains them in a form understandable by humans. We apply
our methodology to analyze connectivity networks between the
assets and discuss our conclusions.
is isolated and can function independently. Interconnectivity
is a fundamental feature of economic systems, including both
traditional financial markets and novel markets. This study aims
to explore the relationships between traditional markets and
commodities trade. We develop a methodology for analyzing the
relationships between five stock market indexes (S&P500, Dow
Jones, BSE, Hang Seng, FTSE) and three commodities (crude
oil, natural gas, gold), based on multimodal publicly available
datasets incorporating structured numerical and unstructured
news and social network data. To find the existence of direc tional associations, we develop an Explainable ML model that
first learns the dependencies between different assets and then
explains them in a form understandable by humans. We apply
our methodology to analyze connectivity networks between the
assets and discuss our conclusions.
Subjects
File(s)![Thumbnail Image]()
Loading...
Name
CIIT2023_paper_20.pdf
Size
8.97 MB
Format
Adobe PDF
Checksum
(MD5):536144a153d8bf63ab6881f2be30c03c