Query expansion based on user friends and interesting_libraries for social book search
Hanjuan Huang, Qiling Zhao · 2017
In traditional book retrieval system, users may not provide accurate information about what they want to query because of the restrictions of expertise,so that they can not get a good result.We provide a new method.First,got set of user's friends and interesting_libraries from his profile.And then,used the probabilistic model to get the similar users whose interested books similar with the books that the user want to search and calculated the similarity.Got the co-occurrence words of original query from the personal digital library of similar users and calculated the co-occurrence rate using the Jaccard index.Finally,calculated recommend of each co-occurrence word.The one with the highest recommend is selected to combine with the original query and generated a new query.We conducted several experiments on a real-dataset collected from LibraryThing.It shows that our method can effectively expand the user original query and improve the accuracy of query.