ARTIFICIAL INTELLIGENCE–BASED ANALYSIS OF DISCOURSES ON SUSTAINABLE AGRICULTURE

Vollständiger Abstract

Worum geht es in dieser Arbeit?

This study aims to investigate how sustainable agriculture is represented in the digital public sphere by examining technology-driven agricultural systems through artificial intelligence–based text analytics. A dataset of 13,354 English posts collected from the X platform during January 2026 was used, with 10,782 posts retained after preprocessing. The methodology integrates text mining, TF-IDF keyword extraction, BERT-based sentiment analysis, and LDA topic modelling. The optimal number of LDA topics was determined using coherence score evaluation to ensure statistical robustness and thematic interpretability. This study integrates text mining, BERT-based sentiment analysis, and LDA topic modelling within a unified artificial intelligence framework to provide a comprehensive analysis of sustainable agriculture discourse on social media. The results show predominantly neutral discourse (65.27%), reflecting informational content, while positive discourse (33.17%) highlights smart farming and innovation. Negative discourse (1.56%) primarily addresses structural challenges. Three dominant themes emerged: climate-oriented sustainability, community-based practices, and technology-driven agriculture, emphasizing the role of digital technologies in shaping sustainable agricultural systems. The findings provide practical insights for policymakers, agricultural stakeholders, and researchers by supporting evidence-based communication strategies and technology-oriented sustainability policies.