Full bibliography

"This study addresses the challenges of multidimensional nonlinearity and multifactorial influences in stock market prediction by proposing a dual-channel deep learning framework based on BERT+BiGRU+Attention (BBA), which integrates investor sentiment analysis with temporal trading data dynamics. Leveraging crawled data from platforms such as East Money Stock Bar, we analyze 30,000 investor comments on Tsingtao Brewery (600600) and corresponding trading records (2010- 2021). Key innovations include: A domain-specific financial sentiment lexicon enhanced by BERT for contextual text vectorization, coupled with BiGRU-based temporal feature extraction and attention-driven keyword focus.Test the rise and fall pages of stocks using stock reviews"