An Association Based Query Expansion Approach Using Linked Data

This work aimed at improving the results of Information Retrieval Systems using DBpedia. Our suggested association based approach consisted on expanding queries with Feedback documents’ indexed terms that were semantically related to query terms using cosine similarity. To evaluate our approach, we used the collection of documents TREC AP88-90. The obtained results showed that our method outperformed the state of art methods, and allowed an improvement of 12.22 percent in Mean Reciprocal Rank at 30 compared to our baseline.