Development and Evaluation of an Automated Methodology for Identifying Framing Patterns in News on Digital Platforms Using Natural Language Processing and Machine Learning

Abstract

The media landscape is saturated with information, making manual news content analysis increasingly impractical. Framing refers to the way media outlets select and emphasize certain aspects of a story, shaping how the public perceives and interprets news events. This thesis presents the development of an automated methodology to identify framing patterns in news articles using Natural Language Processing (NLP) and Machine Learning (ML) techniques. The research is grounded in the creation of a Spanish-language dataset annotated with framing labels, designed to capture a diverse array of news topics and perspectives. By leveraging multi-label classification algorithms, the proposed methodology accurately identifies and categorizes framing patterns, offering insights into how media framing influences public perception.\ The results demonstrate the algorithm's robustness and scalability, with high precision and recall across varied news contexts. This approach not only advances the field of automated media analysis but also contributes to enhancing transparency and objectivity in news reporting. The findings underscore the significant role of media framing in shaping public discourse and highlight the potential of NLP and ML in addressing contemporary challenges in media analysis.