An Abstractive Approach For Text Summarization
Abstractive Summarization creates a shorter and informative version of a text document by extracting important information from the text and generating new sentences using that information. This paper proposes an abstractive technique for text summarization which uses natural language processing for summarization purpose to achieve abstraction. The approach which is discussed in this paper make use of Stanford NLP tools, extraction rules, domain knowledge and sentence generation pattern. The results show that use of natural language processing(NLP) in automatic text summarization can provide a certain level of abstraction which is not possible in statistical approaches. The proposed approach will produce abstractive summaries for a single domain only that is disastrous news and accidents. The findings indicate that based on the parameters information content, reader satisfaction, summary length, grammatical correctness, the proposed approach yields better results as it fulfills the ultimate motive of generating summaries. The findings provide significant analysis that must be updated with the growth of Abstractive Summarization research.
Keywords- component; formatting; style; styling; insert (key words)