Mobile users usually browse web pages on mobile terminals. Many new interesting items occur when the user browses web pages. However, since former methods use the history of past searches to identify users’ interests in order to recommend services based on them, it is difficult to estimate pinpoint and new interests for the users. This paper proposes a method to estimate the hidden interesting items in pinpoint, by the user’s touch operations and pupil reactions. A part of a web page which user looks at is regarded as their interested items when both touch operations and pupil reactions make a response related to their interested items. The methods can deal with users’ interests, because touch operations and pupil reactions show their current interests. Moreover, using both touch operations and pupil reactions improves the precision of the estimation, because they can reduce each noise. Users are able to enjoy the services provided according to their estimated pinpoint and current interests after the estimation of the interested items. When we estimate interested items with the proposed method, we calculated the precision, the recall and the F-measure for every subject. The mean of the precision, the recall and the F-measure are 0.850, 0.534, and 0.603, respectively. In addition, we discuss how to improve the proposed method from the aspects of touch gestures and pupil reactions.
展开▼