Recognizing Textual Parallelisms with Edit Distance and Similarity Degree

Detection of discourse structure is crucial in many text-based applications. This paper presents an original framework for describing textual parallelism which allows us to generalize various discourse phenomena and to propose a unique method to recognize them. With this prospect, we discuss several methods in order to identify the most appropriate one for the problem, and evaluate them based on a manually annotated corpus.

• Type: Conference
• Year: 2006
• Where: EACL
• Authors: Marie Guégan, Nicolas Hernandez
• Keywords: Discourse Structure, EACL 2006, Natural Language Processing, Textual Parallelism, Various Discourse Phenomena