Oxymora are combinations of contradictory or incongruous words and are typically used to attract readers’ attention to certain text. We proposed a method to generate oxymora using an association word corpus and a large-scale N-gram corpus. In the first stage, adjectives were input to the proposed system. Then, antonym pairs were extracted from the N-gram corpus. Candidates were generated using the antonym pairs and the association word corpus. Candidates were finalized or eliminated according to their suitability and attractiveness. Pointwise mutual information (PMI) was employed to exclude grammatically unnatural expressions, thus determining their suitability. PMI, gap of frequency of the oxymoron candidates, and WordNet were used to determine attractiveness. The generated combinations of oxymora indicate the potential effectiveness of the proposed method.
For more details, please click following link below.
Hiroaki Yamane and Masafumi Hagiwara, “Oxymoron generation using an association word corpus and a large-scale N-gram corpus,” Soft Computing, 2015.