Investigating Variations in Adjective Use across Different Text Categories

Adjectives are an informative but understudied linguistic entity with good potentials in sentiment analysis, text classification and automatic genre detection. In this article, we report an investigation of the variations in adjective use across different text categories represented in a sizable corpus. In particular, we report the distribution of adjectives across a range of categories grouped together as academic prose in the British National Corpus. We shall measure inter-category similarity in the use of adjectives and demonstrate with empirical data that adjectives are an effective differentia of text categories or domains, at least in terms of arts and sciences as the two major sub-categories within academic prose. © A. Gelbukh (Ed.)