Support Vector Machine Analysis of Construction Workers’ Automatic Behavior and Visual Attention

Automaticity is a core attribute of skill that is achieved as the attentional requirement gradually diminishes as proficiency increases with practice or repeated execution of a task. Hence, it is a phenomenon that affects workers’ performance both positively (e.g., productivity) and negatively (e.g., accident involvement). Nevertheless, despite its significance, little is known about the effects of automaticity in the construction industry. To address this knowledge gap, this study used eye-tracking technology to examine the effects of automaticity on attention to productivity and safety-related areas of interest (AOIs) during repetitive construction activities. To achieve this research objective, 28 participants were recruited for a simulated roofing experiment. Based on the participants’ attentional distributions, this study employed a support vector machine to distinguish and identify workers who are exhibiting automatic behaviors from those who are not. The findings of this study discovered that although automaticity can improve productivity, it can also cause workers in the construction industry to pay less attention to safety-related AOIs, which puts them at greater risk of safety hazards. Construction managers can use this study’s model to predict automatic behaviors among workers to evaluate productivity and identify at-risk workers.