Content-based navigation in Web-based learning applications

Rapid growth of the Internet make Web-based applications becomes a new means to learn. Authors of Web-based learning applications with multiple resources provide navigation to help users understanding structured idea of learning topics. Since the increasing of resources is such a growing field, navigation map of learning topics would be out of date if it is manually constructed. We define content-based navigation as a sequence list of topics and sub topics hierarchically structured from a collection of documents which is created without relying on human power. Content-based navigation not only provides subjects domain exists in the collection but also offers guidance of relevant items for users who want to learn particular subjects. In this paper, we present a framework for generating content-based navigation from resources applied in Web-based learning applications. We use data mining techniques to extract existing but hidden topics. We do hypergraph partitioning to derive the topics from hyperedges of words correlation which connecting vertices of topics corresponding words. Then we employ agglomerative clustering to merge any overlapping subjects of topics and produce a hierarchical topics list of navigation.