Abstract

An anatomy based summarization method called Topic Summarization and Content Anatomy (TSCAN) was proposed to summarize the content of a temporal topic. TSCAN models the documents as a symmetric block association matrix, in which each block is a portion of a document, and treats each eigenvector of the matrix as a theme embedded in the topic. A temporal similarity (TS) function is applied to generate the event dependencies and context similarity to form an evolution graph of the topic. A unique feature of TSCAN is the introduction of the event segmentation process to extract the semantic construct event before summarization. An ontology database is used for analyzing the main topics of the article using NPL tool and prot

Authors and Affiliations