Automatically Associating Documents with Concept Map Knowledge Models

Concept-map-based knowledge models are widely used for knowledge capture and sharing. When libraries of concept maps are available, those concept maps can provide a useful context for understanding new documents, and the new documents can provide useful annotations to the knowledge models—if the right documents can be associated to the right concept maps. This paper presents ongoing research on a method for classifying documents according the most relevant concept maps in a concept map library, using natural language processing techniques. It presents an algorithm for extracting concept map fragments from a document, introduces a matching procedure, and compares performance of the overall algorithm with a baseline method, with encouraging results.