Product Design Insights from Physiological Signals using CART-Based Analysis of User Experience and Interaction Time

Abstract

The rapid adoption of user-centered design principles has increased the importance of accurately evaluating user satisfaction, as it directly influences product usability and customer retention. A significant proportion of users tend to abandon applications due to poor interaction experiences, making satisfaction assessment a critical requirement in modern systems. Traditionally, user satisfaction has been measured using surveys, questionnaires, and manual feedback collection methods. However, these approaches are often subjective, time-consuming, and incapable of capturing real-time user behavior effectively, leading to inconsistent and less reliable insights. The limitations of existing systems highlight the need for an automated, objective, and data-driven solution that can analyze user interactions more efficiently. To address this gap, the proposed system introduces a Machine Learning (ML)-based framework that utilizes biosensor data to predict user satisfaction. The methodology incorporates baseline models such as Decision Tree (DT) using Classification and Regression Trees (CART), Extra Trees (ET), Linear Regression (LR), and Gradient Boosting (GB) to evaluate performance across both classification and regression tasks. The proposed approach further enhances the CART model by integrating Adaptive Boosting (AB), improving prediction accuracy and robustness. The system is designed to generate dual outputs by predicting interaction duration as a regression task and classifying user satisfaction into High and Medium categories. For real-time deployment and user interaction, the framework is implemented using the Flask web framework, enabling seamless integration between data processing, model execution, and user interface components. The proposed system demonstrates improved performance compared to traditional approaches, offering higher accuracy and reduced prediction errors. This framework provides a scalable, efficient, and real-time solution for user satisfaction analysis, supporting designers and developers in making informed decisions to enhance usability and overall user experience.