AR Extractor : Automatically extracting constraints from android documentation using NLP techniques

Abstract

When developing android apps, it is difficult for the programmers to follow all programming constrains in Android documents. Programming rules are difficult to be manually identified and documented by programmers. As those rules are not so simple, extraction of these kind of rules will be helpful for programmers to avoid new bugs and violations. Sentences with keywords like should, must, important, note, caution, etc., are possible constraints and can carry significant conditional information, that can be helpful for aspiring android developers. In case of developing complex or large android apps, most of the time it is complicated for the programmers to follow all rules since they do not have these rules documented. This paper proposes a novel method called AR-Extractor (Android Rules Extractor) to automatically extract programming constraints using natural language processing techniques(NLP). NLP is a branch of Artificial Intelligence and it involves a set of statistical techniques for identifying parts of speech, entities, sentiment, text classification and other aspects of text. In this project we worked on extracting important text information from Android Developers website by employing the Parts of Speech tagging to understand the relation between each words in a sentence and identify its structure to refine the unstructured text data into structured consolidated format. It can help programmers to reduce bugs, understand programming parameters, improve software maintainability and reliability.