Investigations in single document summarization by extraction method

Due to the rapid growth of the Internet and online information services, huge information is readily made available online. There is no time for anybody to read everything but critical decisions are required to be made quickly. This has led to large-scale research efforts in text summarization. This paper combines contributions by term frequency weight, position weight and commendation weight to form an effective extraction based summary. We have analyzed the significance of each feature and we found that a combination of all three weights results in a better summary as close as human generated summary. We also show that our summarizer outperforms commercially available summarizers like Copernic summarizer and Microsoft Summarizer.