Analyzing the Sentiment of international Trade News in the Context of Sanctions: NLP Approaches

Abstract

The article focuses on exploring the characteristics of natural language processing (NLP) in trade sanctions-related news. Emphasis is placed on identifying lexical and structural features of texts that can affect the quality of automated analysis. The importance of considering context
and cultural differences when evaluating the tone of news is highlighted, along with discussing challenges associated with interpreting economic and political content. An overview of contemporary sentiment analysis methods, including approaches based on machine learning and
neural networks, is presented. Practical aspects of applying these methods to analyze sanction related news, taking into account their specificities and ambiguities, are also discussed.