Document clustering: comparison of ward's clustering and Kohonen network performance / Noor Suriana Abu Bakar and Nurul Nisa Mohd Nasir

Document clustering has been investigated for use in a number of different areas of information retrieval. The aimed of research in the field is to improve efficiency and effectiveness of retrieval. Since the clusters perform best quality, Hierarchical clustering is most commonly used in document clustering. Recently, there exist researches that apply NN in IR. However, research in NN based document clustering still less frequent. Therefore, this study will apply hierarchical based document clustering and NN based document clustering in terms of suggestion supervisor and examiner for thesis. The results from these two techniques will then compare with manual system to find out whether hierarchical based or NN based performed better. The collection of theses will be used and employed the pre-processing including stop word removal and stemming further measure the document similarity before apply the clustering techniques. The result will give some insight whether NN is better for suggestion supervisor and examiner.