Agglomerative Clustering Based Approach to Sentence Ordering for Multi-document Summarization(質問応答・要約,テキスト情報の要約と掲示に関わる自然言語処理シンポジウム及び一般)

概要

論文の詳細を見る
Ordering information is a difficult but an important task for natural language generation applications. A wrong order of information makes a text difficult to understand. When humans order information they use common sense and experience. However, computers are not naturally aware of the order among information. This paper proposes an algorithm that learns orderings from a set of human ordered texts. We cluster sentences into locally coherent blocks and use an agglomerative clustering algorithm which merge these blocks to create a complete summary. Our experimental results show a significant improvement over the existing methods for sentence ordering.
社団法人電子情報通信学会の論文
2006-01-26