Work

Leveraging AI-driven sentiment analysis for real-time user experience feedback in mobile applications

Year: 2025
Type: dissertation
Abstract: Designing and implementing mobile applications that can respond to user emotions in real time and create responsive user interfaces is a novel approach and a focus of attention in the field of user experience improvement. In this thesis, a prototype of an application called Sentimo is introduced that analyses and integrates emotions on the device to adjust user interface elements based on users’ emotional feedback. To process textual feedback in this system, a lightweight version of DistilBERT is used, which classifies the collected data of user emotions as positive or negative and processes the data locally while maintaining the principles of user privacy. A hybrid evaluation was also conducted, which included a survey with 100 participants and receiving qualitative feedback from 42 participants. The survey results show that 91% of participants appreciated the privacy-preserving approach and 83% of users agreed with the emotional understanding and analysis of the system. These findings suggest that the mobile user experience can be improved by using AI-based systems that are aware of emotions and adhere to privacy principles without compromising or questioning user performance or trust. The study also highlights the critical role of AI-based design in real-time applications. Common concerns about data security in emotion-based user interfaces due to data being off-device and cloud-based processing can be addressed by implementing the proposed Sentimo prototype mechanism due to local processing. In addition, due to the positive user evaluation of the system’s ease of use, it is possible to effectively and efficiently embed complex AI behaviour in familiar mobile application environments. The insights from this research provide valuable and practical guidance for future advances in emotion-aware user experience design, especially in privacy-sensitive fields such as education and mental health and personal support programs.
Author: Mohammadian, Leila
Language: English
Open Access status: green
OpenAlex ID: W7111584367
Cites: 37
Cited by:
Related to: 10