DistilBERT and LDA for Understanding Consumer Attitudes Toward Sustainable Products in E-Business

This study found a complex relationship between sustainable practices in the e-business environment and customer purchasing decisions. Through a comprehensive sentiment analysis method, we investigated how sustainability initiatives influence purchasing behavior and brand perception. A novel aspect of this research lies in the direct integration of DistilBERT-based sentiment analysis and Latent Dirichlet Allocation (LDA) topic modeling into sustainability-related product reviews-an emerging approach in e-commerce sustainability analytics. Our findings reveal that while customers highly value environmental considerations, their actual purchasing behavior shows a more complex pattern of adoption and resistance. The results of sentiment analysis on eco-friendly product reviews showed that there were 67.65% negative sentiment, 32.20% positive sentiment, and 0.16% neutral sentiment, indicating that there are significant challenges for sustainable e-business practices despite the growing environmental awareness in the community. Topic modeling acts to identify the key drivers behind both positive and negative responses from customers, generating actionable insights for e-business actors looking for effective sustainability strategies to implement.