Collaborative Kodama Agents with Automated Learning and Adapting for Personalized Web Searching

The primary application domain of Kodama 1 is the World Wide Web and its purpose in this application is to assist users to find desired information. Three different categories of Kodama’s agents are introduced here, Web Page Agents (WPA), Server Agents (SA), and User Interface Agents (UIA). Kodama agents learn and adapt to the User’s Preferences (UP), which may change over time. At the same time, they explore these preferences to get any relevancy with the future queries. The main trust of Kodama research project is an investigation into novel ways of agentifying the Web based on the pre-existing hyper-link structure. These communities of Kodama agents automatically achieve and update their Interpretation Policies (IP) & UP and cooperate with other agents to retrieve distributed relevant information on the Web. We focus in this paper on the implementation and the evaluation on the adaptability of Kodama agents with the UP. This paper proposes a new method for learning the UP directly from user’s interaction with the system and adapting the preferences with user’s responses over the time. The user’s feedback is used by the Kodama to support a credit adaptation mechanism to the IP of the WPA that is responsible for this URL and to adapt the weight and the query fields in user’s query history and bookmark files. In terms of adaptation speed, the proposed methods make Kodama system acts as a PinPoint information retrieval system, converges to the user’s interests and adapts to the sudden change of user’s interests over time.