[2D1-GS-9] Natural language processing, information retrieval: Support technology

[2D1-GS-9-04] A Corpus of Paragraph Titles Given to English Essays

〇Genichiro kikui¹, Shohei Matsuiwa¹ (1. Okayama Prefectural University)

Keywords:Title Generation, automatic summarization

In this paper, we introduce {\em a corpus of paragraph titles}. A paragraph title is a short linguistic expression which indicates or summarizes information of a given paragraph. Paragraph titles are useful representation of a text since they express the organization of the text. In this work, we created a corpus consisting of paragraph titles, composed by humans, for English essay texts. The paper describes statistics of the corpus such as length, overlaps between a paragraph and its title. The paper also evaluates existing title generation algorithms with our corpus.