| The continuous improvement of digital publishing technology provides people a number of e-books.The e-books have many chacteristics,such as digital,convenience,which have been gradually changed people’s reading habits. People can get a lot of knowledge or skills from these electronic books, but faced with too much information so that they have no enough time to read them.And that we extract the rich knowledge from the books will enhance our reading and knowledge service.Therefore, we are hoping to extract the knowledge units rapidly and accurately. Topic research is a branch of text analysis,which can identify the topics and organize the semantic relations by constructing a document structure tree.It can help users search efficiently,conveniently. At this stage, many text analysis method are mostly from the perspective of discourse, paragraphs, and even the full text, ignoring the relationship between topics, hierarchy structure and context information so that can not provide users with satisfactory results. In addition, the existing researches of text analysis are generally inefficient and blindness due to their variety and complexity. Therefore, how to effectively do book topic analysis and organization becomes an urgent problem.This paper mainly includes the following sections:Firstly, a new knowledge organization method for e-books is proposed. Based on the existing document organization theories, we propose a model combined with the characteristics of books, hierarchical topic model and context information to mining the topics of the book.Secondly, this paper design and implement a topic analysis system for e-books using computer technology.Then, experimental results proves that the model is feasibility and practicability. Compared with book content system, it has a high accuracy.Finally, this paper concludes by summarizing the research work, and indicating the future research direction. |
PDF Full Text Request