Twitter Query Expansion via Word2Vec-Urban Dictionary Model
Jason Turner, Mehmed M. Kantardzic · 2018
Query expansion has been a field of interest within the field of information retrieval for quite some time. We propose a novel approach for expanding queries on microblogs, using a word2vec-Urban Dictionary hybrid model built on a prior collection of documents, to expand the search set by adding additional slang terms. In our case, we will focus on the social network site Twitter and tweets related to the topic of marijuana. The result of this approach indicated that we were able to collect several tweets that would not otherwise be collected. We also increased the average degree of the social network of contributors as the expanded list of terms also resulted in marijuana-related tweet contributors who would not have been collected as well.