Assessing and Recommending Brand Features of Electronic Products through Online Reviews and BiLSTM Models

With the advancement of big data and the widespread adoption of e-commerce, there is a growing volume of consumer feedback left after online shopping. Utilizing natural language processing techniques for sentiment analysis on user reviews can unearth consumer perspectives on products. Analyzing sentiment and extracting themes from large-scale product reviews holds immense potential for development and commercial value. In this study, we leverage natural language processing to analyze consumer sentiments and product features in e-commerce. We collected and segmented comments from major e-commerce platforms using Python web scraping tools, focusing on laptop brands. Our analysis, employing Word2vec-BiLSTM models for sentiment analysis and LDA topic modeling for theme extraction, revealed consumer perceptions towards different laptop brands. For example, Apple laptops received praise for memory and performance but were criticized for heat dissipation, while Lenovo was lauded for screen quality but faced concerns about memory and after-sales service. Our findings provide valuable insights for consumers to understand product strengths and for businesses to improve product features and marketing strategies.