From Atoms to the Solar System: Generating Lexical Analogies from Text

Abstract A lexical analogy is two pairs of words (w1, w2) and (w3, w4) such that the relation between w1 and w2 is identical or similar to the relation between w3 and w4. For example, (abbreviation, word) forms a lexical analogy with (abstract, report), because in both cases the former is a shortened version of the latter. Lexical analogies are of theoretic interest because they represent a second order similarity measure: relational similarity. Lexical analogies are also of practical importance in many applications, including textunderstanding and learning ontological relations. This thesis presents a novel system that generates lexical analogies from a corpus of text documents. The system is motivated by a well-established theory of analogymaking, and views lexical analogy generation as a series of three processes: identifying pairs of words that are semantically related, finding clues to characterize their relations, and generating lexical analogies by matching pairs of words with similar relations. The system uses a dependency grammar to characterize semantic relations, and applies machine learning techniques to determine their similarities. Empirical evaluation shows that the system performs remarkably well, generating lexical analogies at a precision of over 90%. iii Acknowledgments