Towards User Interface Adaptation to Users Emotions In a Natural Risks Program
2022.
Abstract
Under conditions of uncertainty, for instance, when people face critical information related to their direct environment, decision-making can be difficult. It can result in systematic bias between the theoretical models and the actual behaviors. Such biases can have significant consequences in domains like risk management.Creating an information chain in integrated risk management governance requires new technologies to collect information, consolidate, integrate, and disseminate it. In this perspective, many users with different behaviors, characteristics, and skills will use these tools. Individual differences urge the need to adapt these tools' user interfaces (UI) to the context of use. Furthermore, emotions as a centerpiece of human characteristics play an essential role in applying various adaptations in such UIs. This thesis approach is based on emotions that can influence user decisions.We first explored the effect of users' emotions in natural hazards preparedness using a simulation video of a hazard. We explored the potential emotional biases and their effect on decision-making. More precisely, running three experiments, the result indicates, in natural hazards, when an individual predicts their emotions intensity, it impacts the preparedness level such that arousing negative emotions leads to an increase in risk perception and potentially adapts the behavior accordingly. We proposed a conceptual model which represents the role of emotional biases on people's preparedness in the context of natural hazards.Second, based on literature studies findings, we understood that UI personalization is based on creating a user model, which allows adapting the interface to the user and their environment. We present a generic problem space to cover various dimensions for UI adaptation. The problem space covers various sources for capturing emotions and various types of emotion models such as Ekman's basic emotions and the Circumplex Model of Affect. Moreover, the problem space considers various dimensions in the temporal aspect of inferring and applying adaptation. Consequently, a generic architecture is proposed based on the problem space that can be used to design adaptive UIs to users' emotions and the adaptation's temporal aspects. A software implementing this architecture can capture run-time users' emotions and adapt the UI at the proper time to users' emotions. The adaptation temporal aspect can improve the tools target outcome. In the final step, a prototype illustrates the architecture to show the approach's feasibility and interest. We chose to create an adaptive interface for a serious game, which enhances public risk perception and preparedness.