Criação de uma ontologia de extração de casos de uso a partir de user stories, utilizando a biblioteca NLTK

Abstract

During the 1970’s the term Software Crisis was used in reference to a Malthusian problem which came to be during the early stages of Software Engineering. During those times, the growth of software demand and complexity increased rapidly, while the growth of software development remained linear. After some time, new development methods were created with the intention of solving the problems related to excessive development times and increase productivity. Some of the methods created were the Structured Programming, SDM (System Development Method), SSADM (Structured Systems Analysis e Design Method), Object Oriented, and Agile. The creation of those methods contributed for the undoing of the Malthusian problem which loomed on the field of Software Engineering. However, the growth of software demand and complexity continues to rapidly increase in a world in which a software is present in every situation and activity. This document proposes the development of a new ontological identification method, main words based and consequently valuation in the phrases, keeping a focus in translation of formatted English natural language texts into software requirements, which can then be used by computer programmers and software engineers. By making use of natural language processing, combined with a chosen set of meta-heuristics, ranking the importance and the frequency of grammatical functions for the development of an ontology to convert well formatted texts into use cases, applied to the Umbrello2 Software, by XMI model. At the end of this document were identified patterns, the importance and behavior of words, grammatical classes. Also shown are statistical data which were used to help characterize, model and comprehend the formation of user-stories from different data sets.