Enhancing Mobile App Reviews with Structured Feedback and NLP-Driven Analysis for Improved Developer Insights and Review Authenticity

The project focuses on enhancing mobile app reviews by addressing the limitations of current review systems, such as unstructured feedback and fake reviews.

The proposed solution introduces a structured review submission system that categorises feedback using predefined tags like “Usability” and “Features,” alongside verification mechanisms, including “Verified Download” and “Verified Purchase” tags, to ensure the authenticity of reviews.

Additionally, advanced Natural Language Processing (NLP) models, such as GPT-4 and RoBERTa, are integrated for automating tag generation and sentiment analysis, providing developers with actionable insights. A static prototype tested by 37 participants showed that the system improves review clarity, authenticity, and trust while offering valuable feedback for developers.

Project Lead

Dr Omar Haggag

Project Team

Prof John Grundy and Prof Rashina Hoda