Agricultural Geographical Indication Brand Management Based on Bi-GRU+CRF Technology and Information Extraction

The application of artificial intelligence technology in the field of information extraction provides a new solution for the management of geographical indication brands in agricultural products. Research combines Bidirectional Gated Recurrent Unit (Bi-GRU) and Conditional Random Fields (CRF) to achieve brand entity recognition, and designs information extraction and diagnostic models for brand management. The results showed that the Bi-GRU+CRF exhibited a good entity recognition accuracy of $94.5 \%$, with better recognition performance than other compared models. It also showed good extraction coverage of $88.14 \%$ and diagnostic accuracy of over $90 \%$ under brand management, with good interpretability. This method provides technical support for the intelligent management of geographical indications for agricultural products and promotes the development of agricultural informatization.