Personalized social query expansion using social annotations

© 2019, Springer-Verlag GmbH Germany, part of Springer Nature. Query expansion is a query pre-processing technique that adds to a given query, terms that are likely to occur in relevant documents in order to improve information retrieval accuracy. A key problem to solve is “how to identify the terms to be added to a query?” While considering social tagging systems as a data source, we propose an approach that selects terms based on (i) the semantic similarity between tags composing a query, (ii) a social proximity between the query and the user for a personalized expansion, and (iii) a strategy for expanding, on the fly, user queries. We demonstrate the effectiveness of our approach by an intensive evaluation on three large public datasets crawled from delicious, Flickr, and CiteULike. We show that the expanded queries built by our method provide more accurate results as compared to the initial queries, by increasing the MAP in a range of 10 to 16% on the three datasets. We also compare our method to three state of the art baselines, and we show that our query expansion method allows significant improvement in the MAP, with a boost in a range between 5 to 18%.