A MINING APPROACH ON DISCOVERING COURSES PATTERN FOR LEARNING PATH CONSTRUCTION

摘要

In recent years, browser has become one of the mostrnpopular tools for searching information on the Internet.rnAlthough a person can conveniently find and downloadrnspecific learning materials to gain fragmented knowledge,rnmost of the materials are imperfect and have no particularrnorder in the content. Therefore, most of the self-directedrnlearners spend most of time in surveying and choosing thernright learning materials collected from the Internet. Thisrnpaper develops a web-based learning support system thatrnharnesses two approaches, the learning path constructingrnapproach and the learning object recommending approach.rnWith collected documents and a learning subject from arnlearner, the system first discovers some candidate coursesrnby using a data mining approach based on the Apriorirnalgorithm. Next, the leaning path constructing approach,rnbased on the Formal Concept Analysis, builds a ConceptrnLattice, using keywords extracted from some selectedrndocuments, to form a relationship hierarchy of all thernconcepts represented by the keywords. It then uses FCArnto further compute mutual relationships among documentsrnto decide a suitable learning path. For a chosen learningrnpath, the support system uses both the preference-basedrnand the correlation-based algorithms for recommendingrnthe most suitable learning objects or documents for eachrnunit of the courses in order to facilitate more efficientrnlearning for the learner. This e-learning support systemrncan be embedded in any information retrieval system forrnsurfers to do more efficient learning on the Internet.