Social book search: Reranking based on document and query expansion with keyword Filtering

Online searching of books have gained astounding popularity worldwide. It has also attracted variety of researchers globally. Searching of books (e.g. Amazon.com, aNobii, LibraryThing etc) with the help of Social metadata(e.g. tags, reviews) and professional metadata (e.g. ISBN Number, Title, Publisher) is gradually becoming a sizzling hot topic under the aegis of Information Retrieval. In this paper, taking Social Book Search as an example, our experiment is divided into three folds: Firstly, in our experiment we expanded the corpus (i.e. Amazon/LibraryThing) by using other book search related catalog sites (e.g. lookupbyisbn.com and Goodreads) for indexing. Secondly, for query expansion, the terms of queries identified in the document, then those terms from the document are appended in our query for re-ranking. Finally, our proposed method extensively evaluated on CLEF 2015 Social Book Search datasets and has better performance compared to other state-of-the-art systems. Recently we got the better performance in CLEF 2016.