Abbreviation Expansion Using a Single Word as Context.

Accurate interpretation of medical abbreviations is vital for clinical decision-making, given their frequent ambiguity in medical texts. We propose a simple Word Sense Disambiguation method using Odds Ratio (OR) of surrounding word tokens to build a knowledge dictionary from a large medical corpus. Abbreviations are expanded using the most informative word identified in this dictionary. Evaluated on ambiguous abbreviations from the NCBI Disease Corpus test set, our OR-based method achieved an accuracy of 0.82, closely approaching SVM (0.87) and BERT (0.89) models. Despite its simplicity, our method highlights its potential as an interpretable and precise alternative for clinical text.