As online users are interacting with many mobile apps under different usage contexts, user needs in an app design process have become a critical issue. Existing studies indicate timely and constructive online reviews from users become extremely crucial for developers to understand user needs and create innovation opportunities. However, discovering and quantifying potential user needs from large amounts of unstructured text is a nontrivial task. In this paper, we propose a domain-oriented deep learning approach that can discover the most critical user needs such as app product new features and bug reports from a large volume of online product reviews. We conduct comprehensive evaluations including quantitative evaluations like F-measure a, and qualitative evaluations such as a case study to ensure the quality of discovered information, specifically, including the number of bug reports and feature requests. Experimental results demonstrate that our proposed supervised model outperforms the baseline models and could find more valuable information such as more important keywords and more coherent topics. Our research has significant managerial implications for app developers, app customers, and app platform providers.
Qiao, Zhilei; Wang, Alan; Abrahams, Alan; and Fan, Weiguo, "Deep Learning-Based User Feedback Classification in Mobile App Reviews" (2020). Proceedings of the 2020 Pre-ICIS SIGDSA Symposium. 8.