Searching for dependencies among concepts in a e-learning system with decision tree

As personalized e-learning service evolves, people have become increasingly demanding on e-learning system to build their knowledge pathway. Semantic web technologies are widely used in describing the data and resources of an e-learning system, such as learners´ profile, learning objects, learning processes, etc. Concept is a fundamental part of learning objects in most semantic web based e-learning system, and the amount of instances of Concept class is enormous and keeps on growing. Prerequisite relation between different concepts is a key attribute for reasoning and personalized recommendation. In the paper, we propose a new decision-tree-based algorithm for detecting the dependencies among concepts. The algorithm transforms learner´s score of a concept into a probabilistic data, and constructing a decision tree to search for the prerequisite concepts accurately and automatically.