1 results listed
The refugee problem is one of the most important
issues worldwide which has gained tremendous media attention
that has influenced public opinion in relation to displaced
populations. This study investigates the application of natural
language processing (NLP) to predict and analyze news articles
on the refugee crisis. The refugee crisis remains an urgent
humanitarian issue with over 117 million forcibly displaced
people by the end of 2024. This study aims to develop an
automated system to predict and analyze news articles on the
refugee crisis using NLP techniques to gain better insights into
migration patterns and highlight key areas for consideration.
Mass migration and asylum-seeking pose major challenges for
host country authorities, non-governmental organizations
(NGOs) and international humanitarian organizations trying to
solve the problems associated with the refugee crisis. Using news
datasets from sources such as The Guardian and Kaggle, the
study refined over 55,000 general category news articles to
extract 6344 refugee-related articles by fine-tuning a Large
Language Model (LLM), “Mistral 7b v0.3”. This study addresses
existing gaps in AI applications by employing LLM to predict key
themes, detect bias, and analyze media narratives on the refugee
crisis. The methodology follows the CRISP-DM framework and
uses pre-processing, prediction and visualization techniques. The
results of this study include highlighting key refugee issues such
as health, shelter, nutrition, security and women’s issues. In
addition, it identifies potential gaps in the treatment of
disadvantaged groups such as LGBTQ+ and disabled people. The
study shows that, LLM outperforms traditional keyword
searches, as three times more relevant articles were extracted
through LLM. The findings of this study give significant details to
strategy makers and NGOs to make more informed decisions on
these important issues. Limitations, such as resource
compatibility and dataset availability, affect the overall results
yet the study highlights the potential of Artificial Intelligence (AI)
in addressing the complexity of global crises and provides a
foundation for future work in multilingual and multimedia
analysis.
International Conference on Advanced Technologies, Computer Engineering and Science
ICATCES
Ahmed Asif
Jagdev Bhogal