Personal Web browsers still don’t support semantic information organization for previously visited Web pages. If such a feature were implemented in a Web browser, then an Internet surfer would be able to easily find important and significant information that was explored in the past. Such an enhancement, in Internet browsing, is being tackled by NTT’s Cyber Solutions Laboratories, which is developing a system that is able to extract and link contextual information from the user surfing experience. This poster paper exposes the system.

The system works by analyzing the content of a Web page and retrieving relevant keywords. Additionally, the system is able to link acquired information from the current Web page in relation to previously judged context. First, the context is defined as “a sequence of Web browsing when many Web pages related to the content of the Web page currently being viewed were viewed intensively and a lot of actions were performed.” The time taken for such a process is called the intensive period, and the system proceeds by collecting action logs, extracting keywords from the current Web page, and extracting past contexts while showing their details.

Through the action logs process, the system collects information about the actions taken by a user (mouse clicks, keyboard strokes, copy actions, print actions, text selection, hypertext transfer protocol (HTTP) headers, Hypertext Markup Language (HTML) source, and so on). Note that the system respects the user’s privacy by implementing file encryption. The second process is keyword extraction from the current Web page, which occurs by applying a variation on the well-known term frequency-inverse document frequency (TFIDF) method [1] used in information retrieval. Basically, TFIDF determines keywords appearing frequently in one text but rarely in others. Also, the system extracts past context by locating intensive periods of previous user interactions, and by finding “the degree of importance of each extracted intensive period to current keywords.” Last but not least, the system displays details and suggestions in an interactive and appealing window.

Finally, by implementing such a system in a Web browser, Internet users can easily remember and retrieve significant information obtained through their Web browsing experience in the past, and in relation to their current browsing activity. This research, whose aim is to enhance Internet user interfaces by preserving conventional information retrieval and enhancing accessibility, is well suited for people working in information science and information retrieval.