Exploring Abbreviation Expansion for Genomic Information Retrieval

Exploring Abbreviation Expansion for Genomic Information Retrieval∗ Nicola Stokes, Yi Li, Lawrence Cavedon and Justin Zobel National ICT Australia, Victoria Research Laboratory Department of Computer Science and Software Engineering The University of Melbourne, Victoria 3010, Australia. {nstokes, yli8, lcavedon, jz}@csse.unimelb.edu.au Abstract Abbreviations are commonly found in- stances of synonymy in Biomedical journal papers. Information retrieval systems that index paragraphs rather than full-text arti- cles are more susceptible to term variation of this kind, since abbreviations are typi- cally only defined once at the beginning of the text. One solution to this problem is to expand the user query automatically with all possible abbreviation instances for each query term. In this paper, we compare the effectiveness of two abbreviation expansion techniques on the TREC 2006 Genomics Track queries and collection. Our results show that for highly ambiguous abbrevia- tions the query collocation effect isn’t strong enough to deter the retrieval of erroneous passages. We conclude that full-text ab- breviation resolution prior to passage index- ing is the most appropriate approach to this problem. 1 Introduction Query expansion is a well-known technique used in Information Retrieval (IR) to address the problem of lexical variation between the query and semantically related terms in relevant documents (Efthimiadis, 1996). While on average query expansion methods, ∗National ICT Australia is funded by the Australian Govern- ment’s Department of Communications, Information Technol- ogy, and the Arts and the Australian Research Council through Backing Australia’s Ability and the ICT Research Centre of Ex- cellence programs. such as relevance feedback (Ruthven and Lalmas, 2003), have been shown to improve retrieval perfor- mance, there are many examples where query effec- tiveness has been significantly downgraded. How- ever, in terminology rich domains where word sense distributions are heavily skewed, query expansion has been shown to have more of a consistent pos- itive effect on retrieval performance. This trend is particularly evident in the passage retrieval task in- vestigated at the TREC (Text REtrieval Conference) Genomics Track (Hersh et al., 2006). In this paper, we investigate the impact of vari- ous expansion term types on passage retrieval ef- fectiveness in the biomedical domain. Our re- sults show that expanding with ontologically related words (synonyms, hypernyms, hyponyms) signif- icantly improves performance; however, abbrevia- tion expansion shows more inconsistent results sim- ilar to those seen in general domain expansion ex- periments. One would expect that the performance of IR systems that index paragraphs rather than full- text articles would greatly benefit from this sort of expansion, since abbreviations are typically only de- fined once in an entire document. We report the results of our investigation on the TREC 2006 Genomic retrieval task. We compare two abbreviation expansion techniques: the first adds abbreviations found in the ADAM database of abbreviations (Zhou et al., 2006a); the second, uses a pseudo relevance feedback strategy to iden- tify query term abbreviations in the full-text docu- ments of an initial set of retrieved passages. Despite the benefit of mutual disambiguation across query terms, referred to as the query term …

Content truncated to 500 words