Knowledge extraction from pointer movements and its application to detect uncertainty

• Published in: Heliyon 2021-01, Vol.7 (1), p.e05873-e05873, Article e05873
• Main Authors: Cepeda, Catia, Dias, Maria Camila, Rindlisbacher, Dina, Gamboa, Hugo, Cheetham, Marcus
• Format: Article
• Language: English
• Subjects: Decision uncertainty; Human-computer interaction; Machine learning; Mouse movement dynamics; On-line survey; Pointer-tracking; Spatio-temporal features
• ISSN: 2405-8440; 2405-8440
• DOI: 10.1016/j.heliyon.2020.e05873

Pointer-tracking methods can capture a real-time trace at high spatio-temporal resolution of users' pointer interactions with a graphical user interface. This trace is potentially valuable for research on human-computer interaction (HCI) and for investigating perceptual, cognitive and affective processes during HCI. However, little research has reported spatio-temporal pointer features for the purpose of tracking pointer movements in on-line surveys. In two studies, we identified a set of pointer features and movement patterns and showed that these can be easily distinguished. In a third study, we explored the feasibility of using patterns of interactive pointer movements, or micro-behaviours, to detect response uncertainty. Using logistic regression and k-fold cross-validation in model training and testing, the uncertainty model achieved an estimated performance accuracy of 81%. These findings suggest that micro-behaviours provide a promising approach toward developing a better understanding of the relationship between the dynamics of pointer movements and underlying perceptual, cognitive and affective psychological mechanisms. Human-computer interaction; Pointer-tracking; Mouse movement dynamics; Decision uncertainty; On-line survey; Spatio-temporal features; Machine learning