Article Preview
TopText Summarization is one of the key fields in the computer science research and its related studies. In this section we discuss the various related work on the summarization and its techniques. The different types of summarization are explained with their limitations over the proposed system.
(Luhn, 1958) is often regarded as the pioneer for automatic text summarization and much of current day research is still find roots in Luhn's approach to summarize text. Luhn proposed that the frequency of words in an article plays a significant role in determining its significance in a summary. The raw text is pre-processed and the sorted in decreasing order of their frequency. (Edmundson, 1969) extended the earlier work done Luhn and Bax- endale by widening the scope for feature extraction. Edmund- son highlighted two new features - importance of cue words and relevance of the title to the summary.
(Hovy, E., & Lin, C. Y. (1999) built upon the work of Edmundson and threw light on the importance of a sentence based on its relative position in the document. (Conroy, J. M., & O'leary, D. P. (2001) for- mutated a text summarizer based on a Hidden Markov Chain. (Erkan, G., & Radev, D. R. (2004)) proposed a graph-based text summarizer LexRank. Each sentence in the document is represented as a node in a graph and the importance of a sentence is proportional to the eigenvector centrality of the node. (Kaikhah, K. (2004) trained a neural network using features selected from a document to generate a summary.