Automatic classification and indexing technology for literature based on natural language processing in intelligent libraries

With the digital transformation and intelligent upgrading of library resources, the traditional methods of literature classification and indexing can no longer meet the needs of efficient and accurate information organization. In view of this, this study proposes a new method of automatic literature classification and indexing which integrates NLP (Natural Language Processing) technology. The research aims to improve the information processing efficiency and service quality of the library. In terms of technical realization, this paper innovatively designs an automatic literature classification algorithm based on Bert (Bidirectional Encoder Representations from Transformers) model and an automatic literature indexing algorithm based on keyword extraction. By pre-training and fine-tuning BERT model, the in-depth understanding and accurate classification of literature content are realized. By analyzing the lexical frequency, part of speech and location information in the literature, the keywords representing the theme of the literature are successfully extracted, and the automatic indexing of the literature is realized. The experimental data show that the accuracy of the classification model has reached a high standard, and it can effectively distinguish literatures in different disciplines. The indexing model also performs well, and the extracted keywords are highly consistent with the manual indexing results. This technology is helpful to improve the efficiency and accuracy of literature processing and provide users with more personalized and accurate information services.