Sequential pattern based multi document summarization — An exploratory approach

摘要

Sequential Pattern Mining which aims to discover all frequent sequences of itemsets (patterns) from a large data collection has been applied in the Text Mining domain such as Text Categorization and Pattern Identification. However, in the area of Document Summarization the effort is still considered as green and exploratory. In the real world, a sentence is more than just a collection of un-ordered sequence of words, where each sentence carries their own meaning. By discovering these textual patterns is essential since the patterns can describe the text, by preserving the sequential order of the words in the document. Thus, the motivation here is to investigate the feasibility to develop a Sequential Pattern-based Summarizer model near future in order to reduce redundancy information from multiple text resources; at same time preserving the meaning of the original text document using the Semantic similarity approach. This paper reviewed some of the existing techniques in the area of multiple document summarizations to better understand the gap and issues underlying this area. By incorporating the semantic knowledge of sentences in the multiple documents is hoped to assist and alleviate the long-winding process for non-subject expert researches in trying to find the similarities and correlation between text resources.