Enhancing-Product-Design-through-AI-Driven-Sentiment-Analysis-of-Amazon-Reviews-using-BERT

Understanding customer emotions and preferences is paramount for success in the dynamic product design landscape. This paper presents a study leveraging an innovative Natural Language Processing (NLP) model, Aspect-Based Sentiment Analysis (ASBA), using a fine-tuned Bidirectional Representation for Transformers (BERT) model to predict customer emotions. The model is trained on a dataset of 90,000 labeled synthetic reviews, achieving an impressive validation accuracy of 98.8%. Manual content analysis is employed to validate its performance. Our research focuses on different product categories, harnessing web scraping techniques to extract vast customer review data from Amazon.com. Traditionally, manual qualitative analysis demanded significant human effort to predict customer emotions. However, our NLP model automates this process, capturing consumer sentiments within minutes and eliminating the need for laborious human intervention. Comparing the ABSA approach with manual content analysis showcases the efficiency of automation in deciphering customer emotions. By detecting aspects and sentiments of input data using the pre-trained BERT model, our study demonstrates its capability to comprehend and analyze customer reviews effectively. These findings can empower product designers and research developers with data-driven insights to shape exceptional products that resonate with customer expectations. This work embodies embracing cutting-edge technology to propel product design and development, unlocking sentiments hidden within online reviews, and creating meaningful, impactful products aligned with customer needs.

Note The Classification Model and Tokenizer have to be saved and then call them in the predictioon pipeline.