Abstract：A Study on the Issues Related to Building a Library Information System Based on Deep Learning AbstractHigh-performance information systems are essential for libraries to carry out their work efficientlyRecentlya digital library environment has been created that allows electronic resources such as e-bookse-journaland the Web to be linked to SNSsmartphones and tablet PCsThe environment available anytime and anywhere has increased the use of digital librariesand the amount and scope of information search has also increased or expanded dramaticallyThese changes call for improved accuracy in the search for internal or external information in the library information systemThis study designs library information retrieval system from the perspective of QA systemand builds a model based on text dataThrough thisfinallywe present a study comparing performance with various algorithms based on perplexityRetrieval and IndexingThis article mainly describes OPACa information retrieval system that allows users to access information in the electronic libraryand emphasizes the progress of its search and search functionThis article mentioned that the third-generation OPAC system improved the search functionprovided automatic guidance and search assistance to search textor expanded accessability by connecting to various databasesThe most important thing is that the search based on the context and natural language expression adds automatic error correctionThis part involves natural language processingNLPbecause it needs to understand and generate natural language to improve the search and retrieval function