Engineering Web resource summaries using Pointwise mutual information (PMI)-based for web document summarization

Internet is an ocean of information which is growing everyday. It is impossible for any human being to read the whole information. It is very difficult to decide which document is relevant. If someone gets summary of the document that is the easiest way to decide. Therefore, there is a need to create automatic text summarization systems which generate a precise summary. In this paper, a study of summaries generated using PMI based web document summarization technique, which creates summary on ranking of sentences on the basis of Total PointWise Mutual Information of sentences. The data is gathered basically from documents of engineering domain, for text summarization. The results show that the method outperforms the other techniques by exhibiting the best results for the closest mean score and generating a good quality summary of sentences of different lengths. It also depicts that the proposed approach has minimum error compared to other machine learning approaches.