“Pragma SUM: Key Word Use in Automatic Summarization

As access to the Internet broadens and with the advent of tools that allow people to create content, the amount information to which we have access grows exponentially. Texts written about various subjects and by countless authors are produced every day. It is impossible to absorb all the information available or to select the most adequate piece of information for a certain interest or public. Automatic text summarization, in addition to presenting a text in condensed form, can simplify it, thus generating an alternative for saving time and widening access to contained information for many different types of readers. The automatic summarizers that currently exist in literature do not present personalization methods for each type of reader and, consequently, generate results that have limited precision. This article aims to use the automatic text summarizer PragmaSUM in educational texts with new summarization techniques using keywords. Personalization methods using keywords seek to increase precision and improve the performance of PragmaSUM and its summaries. In order to achieve that, a corpus was formed exclusively by scientific articles in the field of education in order to conduct tests and comparisons between different summarizers and summarization methods. The summarizers’ performance was measured by the Recall, Precision and F-Measure metrics, all of which are present in the ROUGE tool. The results point towards improved performance when employing keywords in summarization with PragmaSUM, which suggests the importance of choosing keywords adequately for classifying the content of the source text.