Automatic Text Summarization Using Term Frequency, Luhn's Heuristic, and Cosine Similarity Approaches

People these days are quite busy with their work and other activities. They do not have enough time to fully read newspapers. But they need to keep themselves informed about current events. As a result, text summarization can help us learn about events taking place throughout the world quickly and conserve our precious time. Hence, automatic text summarization is a mechanism for condensing textual information into a brief and understandable piece of knowledge that gives readers an idea of gist from a large passage within a very short time because literature and texts written in regional languages are frequently much long and usually there are no summaries that correlate to them that explain the main point. Although there are a number of research works by many authors, but in our paper, we have attempted to find the summary of paragraphs by three extractive text summarization techniques, namely, term frequency based algorithm, Luhn's algorithm and cosine similarity method. We have implemented our methods using python, and show the sample output for sample input. Moreover, we see that the cosine similarity approach shows lower execution time, and as a human reader, we notice that cosine similarity is the good one to summarize a paragraph.