購買行動における視覚支援システムのための視線追跡による機械学習を用いた意図の分類

概要

Eye tracking offers insight into purchasing behavior, but prior work mainly targets single intents. Real behavior involves sequential shifts—exploration, comparison, selection—largely unaddressed. We sequentially estimate intent from eye-tracking data using machine learning. Features include fixation duration, gaze travel distance, variability, and re-fixation rate. An SVM classifier achieved intent estimation across phases, demonstrating the feasibility of intent-aware support systems.

田村 雄介

准教授