A Smart Framework for Optimizing User Feedback Prioritization in Application Development

In mobile app development, user reviews are a significant source of requirements. Users frequently report bugs, request new features, or suggest enhancements. Mobile app vendors aim to maximize user satisfaction by addressing these continuous comments and requests as early as possible. Typically, they prioritize delivering the most promising features, extracted from user reviews, in early releases while deferring fewer promising ones to later stages. However, due to the massive volume of reviews, redundancy, and conflicts among them, manually extracting requirements is inefficient and often challenging, making requirement prioritization even more difficult. Therefore, automating this process is essential. This paper presents a conceptual framework for the requirements prioritization process's automation, continuity, and scalability. The proposed framework follows a hybrid approach that integrates multiple advanced techniques: generative artificial intelligence, active learning, ontologies, and optimization algorithms. Generative artificial intelligence enables the identification of important patterns and the automatic extraction of requirements and their properties, which aids in assessing properties to determine requirement priorities. The generative artificial intelligence framework integrates with an active learning system to improve annotation efficiency. Ontologies help comprehend relationships, properties, and dependencies among requirements, aligning them with domain-specific knowledge. Optimization methods playa crucial role in the requirements prioritization process by computing the weights of various properties to identify the most effective combination of requirements and determine the optimal order for implementation. Consequently, this research presents a smart theoretical framework for enhancing user-driven maintenance and development of mobile applications. Researchers are tackling several critical challenges that remain unresolved in the field, with future directions focusing on evaluating its applicability and effectiveness in real-world scenarios.