National Technical University of Athens, 9 Heroon Politechniou Str, 15773, Athens, Greece.

---

Abstract—Abbreviations are widely used in clinical studies. Identifying their meaning is an essential step for further processing of the provided data. In this work, a novel system developed for detecting the meaning and context of the abbreviations specified in the whole corpus of clinicaltrials.gov studies is presented in detail. In the background, innovative algorithms and techniques were used for both acronym and non-acronym type abbreviation recognition purposes, while their evaluation was performed based on an abbreviations annotated corpus of clinical studies, which was purposely developed. The outcome of this work is a repository consisting of approximately 28 thousand abbreviations, while for each of them its possible senses along with their context have been recorded, so that it can be utilized for inferring the meaning of abbreviations met in existing and future documents in this domain. Further analysis of the recorded data indicated, amongst others, that there are considerable differences among abbreviations and their expansions or senses depending on the scientific domain in which they are used. Also, the abbreviations consisting of three and especially two characters are already highly ambiguous and their possible senses are going to be significantly increased in the following years.

Index Terms—Abbreviation recognition, clinical studies, sense repository.

[PDF]

Cite: Efthymios Chondrogiannis, Efstathios Karanastasis, Vassiliki Andronikou, Theodora Varvarigou, "Building a Repository for Inferring the Meaning of Abbreviations Used in Clinical Studies," Journal of Computers vol. 12, no. 1, pp. 76-88, 2017.