Unsupervised knowledge navigation: Reconstructing the hypermedia structure of instructional materials on world wide web

There are many different terms when learner surf on web. When learner is reading an instructional material, he/she may not understand the meaning of specific term/keyword in a moment easily. At this time, the learner will need to find out the definitions or references about the keyword. However, unfortunately, the learner will not be able to find out what he/ she is looking for in the most of time. It is because of the instruction designers of the learning materials who had never thought that would be a problem for learners. Therefore, it will be excellent, if the associated documents with the keyword which is the learner looking for could be retrieved automatically, and furthermore, the original document structure could be reconstructed to a more suitable for learning and reading in real-time. In this paper, the graph theorem and the data mining technique are applying to produce an association lattice of keywords and design a searching algorithm of reconstructing the necessary instructional materials on the website automatically.