Research On Summarization Of Book Chapter Based On Reading Behavior

With the development of social network and mobile Internet, the traditional method of digital reading hasn’t met the reading needs of readers. As a new field of digital reading, the social reading has been paid more and more attention. When the readers are reading books, a mass of valuable information, which not only implies the reading habit of the readers, but also implies the value of the book itself, is hided in the data of reading behavior. The traditional digital reading can only provide basic reading services for readers while the social reading can grasp the reading habit of reader better with the help of the analysis of reading behavior on the basis of traditional digital reading service and can provide better reading content and reading experience for readers.As a subsystem of Engineering Science and Education Library Knowledge Service System, the Reading and Noting System is a reading and learning platform that combine the technology of social reading and crowdsourcing. The system can provide better service for readers by collecting, analyzing and modeling on the data of reading behavior of the reader when they are browsing and reading books.The study of our work is mainly on generating the summarization of the specific book chapter and evaluating the importance of the page in a book in consideration of the reading behavior of readers. In the scoring system of evaluating the importance of pages, we use the concept of depth of reading to divide the book reading into four different levels. In the implementation of the algorithm that generates summarization of book chapter, we spread the weight of sentence by the method called ranking on data manifolds. And at last, the summarization can be generated by the method data linear reconstruction.