A probabilistic approach for eye-tracking based process tracing in catalog browsing

Abstract

Eye movements are an important cue to understand consumer decision processes. Findings from existing studies suggest that the consumer decision process consists of a few different browsing states such as screening and evaluation. This study proposes a hidden Markov-based gaze model to reveal the characteristics and temporal changes of browsing states in catalog browsing situations. Unlike previous models that employ a heuristic rule-based approach, our model learns the browsing states in a bottom-up manner. Our model employs information about how often a decision maker looks at a selected item (the item finally selected by a decision maker) to identify the browsing states. We evaluated our model using eye tracking data in digital catalog browsing and confirmed our model can split decision process into meaningful browsing states. Finally, we propose an estimation method of browsing states that does not require the information of the selected item for applications such as an interactive decision support.

A two-step approach for interest estimation from gaze behavior in digital catalog browsing

“…User decision making consists of several phases (Russo & Leclerc, 1994;Schaffer, Kawashima, & Matsuyama, 2016) such as browsing a catalog to acquire information (screening phase) and comparing items to evaluate them (comparison phase). Note that the comparison phase is expected to contain more clues to a user's values than the screening phase.…”

“…Note that the comparison phase is expected to contain more clues to a user's values than the screening phase. To identify the decision phases, a shortterm (segment-wise) analysis of gaze region sequences has been proposed (Ishikawa, Kawashima, & Matsuyama, 2015;Schaffer et al, 2016). A tri-gram of gaze region sequences is considered to be a unit of analysis to extract layout-related gaze features for classifying the decision phases (browsing states).…”

“…The temporal patterns of gaze targets convey useful information for estimating user interest during decision making. For example, if a user "re-fixates" on a target, he or she is probably specifically focusing on the target and comparing it with the alternatives (Schaffer et al, 2016). Although these temporal patterns are indirectly considered in the AOF detection step to identify bias in the user's gaze behavior in a short-term window, explicit modeling of gaze-transition patterns needs to be incorporated for a natural interpretation of user gaze behavior.…”

“…However, the layouts in actual catalog content are not randomized but rather structured (e.g., similar items are arranged more closely together). This affects content browsing and is actually useful for analyzing browsing states in a deci-sion-making process (Schaffer et al, 2016). Therefore, the information of content design needs to be considered for interest estimation.…”