Requirements Engineering in the World of Apps

Abstract

We are living in a world that is rapidly growing digitally, and businesses are becoming increasingly dependent on information. In order to adapt to this growth, and gain competitive advantage, businesses seek new innovative approaches and communication channels to extract new software requirements from online user data. Previous and current research mainly focus their efforts on exploring social media and mobile application platforms for requirements discovery and extraction. In this research, we focus on exploring requirement extraction for desktop applications from technical user forums. We build a prototype tool for scraping a user forum, processing the textual data with NLP tools, and visualizing the resulting data using a visual analytics tool. The classification accuracy on the data set is between 60-90% while the recall between 80-90%. The Naïve Bayes Classifier outperformed other binary classifiers for the data of this research domain. We conducted experiments and interviewed experts to evaluate the perceived usefulness of our prototype. Results show positive feedback on our prototype as effective and efficient tool to support product managers discovering requirements and new markets.