Name Entity Recognition of E-Commerce Product Reviews Based on BERT

The rapid growth of online shopping highlights the importance of entity recognition in e-commerce reviews, a key task for text mining and knowledge graph construction. Due to the lack of public datasets, this study scrapes Apple Watch reviews from JD, processes the data, and builds an entity recognition datasets via semi-automated and manual labeling. Experiments compare CRF, BiGRU, BiGRU-CRF and the proposed BERT-BiGRU-CRF model, with the latter achieving precision, recall, and F1 scores of 83.26 %, 80.85 %, and 81.44 %, outperforming baseline models.