Capturing User's Interest from Human-Computer Interaction Logging

Mining user’s expectations and interests has become the focus of many Internet-based application providers, such as those operating in the areas of social networks, search engines, e-commerce, and so forth. This is often accomplished by means of explicit feedbacks requested to end-users, which might yield distorted results due to the intrusive nature of this kind of approach. Thus, it would be desirable using implicit feedbacks, provide that they faithfully reflect user’s habits and expectations. In this paper we propose an approach to capture user’s feedbacks from their interaction actions while processing a document, with par- ticular emphasis on web documents. To this end, we propose a new model to interpret mouse cursor actions, such as scrolling, movement, text selection, while reading web documents, aiming to infer a relevance value indicating how the user found the document useful for his/her purposes. We have implemented the proposed model through light-weight components, which can be easily installed within major web browsers as a plug-in. The components log mouse cursor actions that we have used as experimental data in order to validate the proposed model. The experimental results show that the proposed model is able to predict user feedbacks with an acceptable level of accuracy.