Mechanism of cognitive style effects on hazard recognition of construction workers: A machine learning-aided approach

BackgroundHazard recognition is a critical skill for construction workers and primarily consists of two stages: visual search and cognitive processes. Cognitive style, categorized as field-dependent (FD) and field-independent (FI), significantly influences this process. FD individuals rely more on external cues and contextual information, while FI individuals are more detail-oriented and analytical. However, the mechanism by which the cognitive style affects this efficiency remains unclear.ObjectiveThis study aimed to clarify the impact mechanisms of FD and FI on hazard recognition and to validate these mechanisms using machine learning.MethodsThis study used eye-tracking and electroencephalography technology to quantify these two stages. The experiment was conducted with hazard recognition as the task and involved 40 participants divided into FD and FI groups. Subsequently, statistical methods were used to compare identification performance, eye-tracking features, and event-related potentials among participants with different cognitive styles. Finally, we applied multiple machine learning algorithms to further verify the impact of the cognitive style on hazard recognition.ResultsFI individuals were faster in hazard recognition than FD individuals, whereas FD individuals were more accurate than FI individuals. Cognitive style mainly affects hazard recognition by affecting the visual search phase.ConclusionsBased on machine learning and multimodal data analysis, this study provides a new perspective for understanding the relationship between cognitive style and hazard recognition. The findings offer a scientific basis for assessing workers' hazard recognition capabilities and implementing personalized intervention management.