Sentiment Analysis on Biodegradable Trash Bags Using a Fine-Tuned BERT Model

Biodegradable trash bags provide an eco-friendly alternative to traditional plastic bags, but their effectiveness and consumer perception remain under-explored. This study aims to analyze customer sentiment toward biodegradable trash bags using customer reviews. 949 reviews were collected and subjected to comprehensive data preprocessing techniques, including text cleaning, tokenization, stopword removal, and lemmatization. Initial sentiment analysis was conducted using the pre-trained BERT model "distilbert-base-uncased-finetuned-sst-2-english". To enhance sentiment classification accuracy, the model was fine-tuned with domain-specific data. Reviews with ratings below 4 were labeled as negative, while those with ratings of 4 and above were labeled as positive. The fine-tuned BERT model achieved an 85.37% accuracy in sentiment classification, improving from the baseline performance of the pre-trained model. Our sentiment analysis on 949 Amazon reviews revealed that 675 reviews (71%) were positive, while 274 reviews (29%) were negative. These findings provide valuable insights for manufacturers, marketers, and policymakers, helping them understand consumer concerns and preferences better. This, in turn, can guide product improvements and more effective marketing strategies to promote the adoption of biodegradable trash bags.