3.1 Knowledge Representation
Knowledge in a domain is represented by logically partitioning different documents based on certain proper- ties so as to place similar documents in same cluster. Initially there was a hard clustering approach in which the documents were associated with one cluster exclusively but gradually it was found that one document may have some degree of association to one cluster while some degree of association to other clusters and hence fuzzy c-means clustering is used here to logically partition documents into various clusters, so that similar documents logically belongs to the same cluster.