Scholarlink

With the increasing use of mHealth applications for wellbeing, support, and monitoring, the Persuasive Systems Design (PSD) model has also become more prevalent in digital health interventions. However, understanding how users perceive and respond to these persuasive strategies remains limited. Harnessing machine learning models, this study proposes a novel methodological approach to examine the user sentiments and their link with the PSD features as expressed in the app reviews. Sentiment analysis was performed on reviews extracted from 10 applications in the Google Play Store, followed by automatic classification of reviews into PSD categories using a transformer. The results revealed a polarised perception across categories, with primary task support predominantly receiving positive sentiment, while system credibility support attracting more negative sentiment. These findings suggest that the implementation of the primary task support features aligns well with users’ needs, and their positive perception can further enhance user engagement. However, the developers need to reconsider the way the credibility support features are embedded since it may be counterintuitive and reduce users’ trust.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Understanding User Perspectives on Persuasive Systems Design Principles in mHealth Applications Using Deep Learning Techniques

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Sudarshan Khadka,
  • Elena Vlahu-Gjorgievska

摘要

With the increasing use of mHealth applications for wellbeing, support, and monitoring, the Persuasive Systems Design (PSD) model has also become more prevalent in digital health interventions. However, understanding how users perceive and respond to these persuasive strategies remains limited. Harnessing machine learning models, this study proposes a novel methodological approach to examine the user sentiments and their link with the PSD features as expressed in the app reviews. Sentiment analysis was performed on reviews extracted from 10 applications in the Google Play Store, followed by automatic classification of reviews into PSD categories using a transformer. The results revealed a polarised perception across categories, with primary task support predominantly receiving positive sentiment, while system credibility support attracting more negative sentiment. These findings suggest that the implementation of the primary task support features aligns well with users’ needs, and their positive perception can further enhance user engagement. However, the developers need to reconsider the way the credibility support features are embedded since it may be counterintuitive and reduce users’ trust.