Intelligent agent for knowledge tree administration

Knowledge tree is a list of pointers to applicable documents WWW pages or other information resources together with the tools that are needed for interactive construction of trees by users. It is in fact virtual multimedia textbook and as such represents appreciated high quality supplement to the existing learning materials. One of the important characteristics of virtual textbooks is. their ability to change dynamically and to comprise new learning materials as soon as they appear in the Web or anywhere in the cyberspace. For these reason, knowledge trees need to be maintained regularly to avoid unpleasant surprises. We upgraded knowledge trees with an intelligent agent that can automatically verify the validity of links; when changes are detected in the document or referenced document is not found after several attempts, the link is marked and administrator can take appropriate measures. But the agent can take even more active role in the tree maintenance process: it can be upgraded to look for new materials to be added to knowledge trees. Documents gathered in knowledge trees, or more precisely links in the documents, represent starting point for our agent when we activate it to extend the existing knowledge tree. The agent checks whether the documents pointed to by these links tit into knowledge tree. There are some parameters to be set before its exploration starts. One of the most important parameters is depth of searching. Agent can follow certain link and recursively links from the document, pointed to by the original link. The value of this parameter have to be selected carefully as the number of links can grow exponentially. Other parameters that can be set for the agent are specifications about the types of links not to be followed during a search for new materials, specifications of links that should always be followed and intermediate cases where decisions are taken according to some specific settings.