Too Long-Didn’t Read: A Practical Web Based Approach towards Text Summarization

摘要

In today’s digital epoch, people share and read a motley of never ending electronic information, thus either a lot of time is wasted in deciphering all this information, or only a tiny amount of it is actually read. Therefore, it is imperative to contrive a generic text summarization technique. In this paper, we propose a web based and domain independent automatic text summarization method. The method focuses on generating an arbitrary length summary by extracting and assigning scores to semantically important information from the document, by analyzing term frequencies and tagging certain parts of speech like proper nouns and signal words. Another important characteristic of our approach is that it also takes font semantics of the text (like headings and emphasized texts) into consideration while scoring different entities of the document.