Research On Chinese Book Classification Based On LSTM Model

Book classification is one of the most important and complex tasks in library and information service.It has long been done by book authors or book catalogers manually.However,with the advent of the information age,the number of books is increasing,and the field of knowledge involved in books is more and more broad.So,the work of manual book classification seems inadequate.Under the background of developing digital libraries,the introduction of computer automation technology into the field of book classification,to realize automatic classification of books,has become a hot topic in the field of library and information.This is of great significance to realize the digitization and sharing of book resources.At present,there have been some research results in the field of automatic classification of Chinese books.In the early years,the classification methods based on knowledge engineering were mainly adopted.In recent years,the classification methods based on machine learning began to rise.However,taking a wide view of the current research results,these classification systems need to be constantly updated and maintained.When the new research fields and new knowledge appear,the knowledge-based classification methods require experts to develop new rules to improve the knowledge base,and the classification methods based on traditional machine learning need to update the feature set and re-train the models.This will undoubtedly bring some difficulties to the implementation of the classification systems.In recent years,with the rise of deep learning,artificial intelligence has entered a new research stage.Recurrent neural networks,convolution neural networks and some other models have been proved to have advantages in various fields,including the field of natural language processing.In this paper,the LSTM model is introduced into the field of automatic classification of Chinese books,and a classification system with practical application value is proposed.The main work of this paper includes the following aspects:(1)First of all,this paper investigates the research status of related fields,including the research status of text classification at home and abroad,and the research situation of Chinese book classification.The article analyzes the advantages of the current research results and their existing problems.(2)The technical theories of text classification are studied in this paper.Book classification is a sub-domain of text classification.Through the analysis of the related technologies of text classification,we can understand the technical points involved in the process of book classification,thus laying a foundation for proposing our new Chinese book classification method.(3)This paper makes a deep study of deep learning,including its background,development,technical points and so on.In particular,the LSTM model is introduced and analyzed in detail,which lays the groundwork for the subsequent design of the Chinese book classification system.(4)This paper designs a Chinese book classification system based on the LSTM model,and designs three experiments:the experiment for the coarse-grained book classification,to explore the impact of different measures on the classification results;the experiment for fine-grained book classification,to compare the advantages and disadvantages of direct classification and layer-by-layer classification;the experiment for coarse-grained book multi-classification,to explore the feasibility of multi-classification.(5)Based on the analysis of the experimental results,the model in this paper has a good performance in Chinese book classification.This model also overcomes some shortcomings of previous book classification systems.Therefore,the Chinese book classification system based on LSTM model proposed in this paper has practical application value.