Proposal of a Novel Book Content Search Method for Web-Based EBook Libraries

Due to high demand of eBooks, libraries of many academic institutions have started web-based eBook library (eBook library, for short) services. Recently many users have begun to search information in eBooks. In order to serve this requirement, many eBook libraries provide a search interface that employs Full-Text Search methodology. However, the search interface has the following two drawbacks. First, it can't perform book content search across eBooks of its library by a single query. Second, although the search interface can indicate the pages containing a keyword inputted by the user, the user has to read page by page until the required content is found. The objective of this research is to propose a novel book content search method which can solve the above two drawbacks. The proposed search method focuses on searching eBooks' contents that explain a keyword given by a user. In order to realize this search function, we utilize both metadata and BoBIs (Back-of-the book Indexes) of the eBooks. To the best of our knowledge, there exists no previous work proposing this search method. We present a prototype of this search method applying to Maruzen eBook library which is used in our university and many universities in Japan.