Requirements Classification and Identification Approach for E-Collaboration Systems.

In the background of global economic integration and the pandemic era, e-collaboration systems have become the main tools for people to learn and work. E-collaboration systems in the Internet era face rapid iteration requires, it’s important to research how to quickly access the requirements of e-collaboration systems to help developers better update and iterate on them and improve the user experience. As the number of user reviews in the application market is large, covering a wide range, a large number of requirements are contained in it, therefore, this paper starts from two dimensions of functional requirements and non-functional requirements, and proposes 15 requirements categories for the collaboration characteristics and timely characteristics of the e-collaboration system. First crawl the collaboration system user reviews as a dataset, use TextRank ranking algorithm and TextCNN model to classify them, and finally by building a network of co-occurrence relationships between words to feed the users most concerned in a certain classification to the developers. Our method has proven to be excellent in terms of Precision, Recall and F1-measure, and can help developers to obtain requirements from large amount of review data.