Automated multilingual climate change sentiment analysis using hybrid AI model

Abstract

Climate change is among the most widely discussed topics on social media, blogs and online news, generating extensive public debate. This study presents an automated sentiment analysis framework that utilises Natural Language Processing (NLP) and Machine Learning (ML) to understand how people perceive climate change. Multilingual, cross-platform textual data were gathered, pre-processed and transformed into numerical features using TF-IDF and contextual word embeddings. Traditional classifiers - Support Vector Machine (SVM), Random Forest (RF) and Long Short-Term Memory (LSTM) - were evaluated, with LSTM achieving the best results in capturing contextual dependencies. To further improve sentiment detection and event-aware trend analysis, four hybrid algorithms are proposed: Adaptive Contextual Polarity Network (ACPN), Hierarchical Emotion Fusion Model (HEFM), Dynamic Event-Aware Sentiment Mapper (DEASM) and Cross-Lingual Semantic Alignment Classifier (CLSAC). The system recognises changing public attitudes in response to environmental events and policies, offering researchers and policymakers deeper insight into global climate discourse.