An analysis of hierarchical clustering and neural network clustering for suggestion supervisors and examiners

Abstract

Document clustering has been investigated for use in a number of different areas of information retrieval. This study applies hierarchical based document clustering and neural network based document clustering to suggest supervisors and examiners for thesis. The results of both techniques were compared to the expert survey. The collection of 206 theses was used and employed the pre-processed using stopword removal and stemming. Inter document similarity were measured using Euclidean distance before clustering techniques were applied. The results show that Ward’s algorithm is better for suggestion supervisor and examiner compared to Kohonen network.