Sentiment-Enhanced Text Mining of Social Media Data to Predict FinTech Consumer Behaviour and Digital Economy Adoption Rates

The rapid evolution of Financial Technology (FinTech) has reshaped global financial ecosystems, enabled greater financial inclusion while introduced new behavioral dynamics. This study proposes a sentiment-enhanced text mining framework to analyze public perceptions of FinTech and predict consumer adoption behavior in the digital economy. Using over 98,000 multilingual social media posts, the framework integrates Transformer-based sentiment analysis (BERT/IndoBERT), Latent Dirichlet Allocation (LDA) topic modeling, and Random Forest regression to quantify the relationship between sentiment polarity, thematic focus, and adoption behavior. The results indicate that positive sentiment toward usability and innovation strongly correlates with higher adoption levels, while negative sentiment surrounding data privacy and security remains a major deterrent. The model achieved an explanatory power of R² = 0.87, validating the predictive capability of sentiment-driven features. Moreover, the use of SHapley Additive exPlanations (SHAP) enhances interpretability, allowing transparent identification of influential variables. This research contributes a novel hybrid framework that bridges emotional analytics and behavioral modeling, offering policymakers and FinTech providers actionable insights into consumer trust, technology perception, and participation in the digital economy.