Improvisation of the Document Summarization by Combining the IR Techniques with “Code-Quantity Memory and Attention” Linguistic Principles

In this paper we have given a new statistical approach for automatic text summarization by combing the Information Retrieval (IR) techniques with the linguistic principles code quantity, memory and attention to get the relevant sentences. In our approach we are distilling the most important information from the source to get the actual concept in the abridged form. We have initially removed the redundancy of the input document by using the Synonymous Cosine Similarity and we ranked the sentences based on the linguistic principles code quantity, memory and attention. Moreover, this method has been run over the test data, obtaining satisfactory results in the evaluation when compared with the MS Word Automatic Summarizer with respect to the human judgment.