Pui Lam Leung (Statistics, CUHK)

Title: Unidimensional scaling, Clustering and Data Mining

Abstract: Unidimensional scaling (UDS) is first briefly introduced. In 
particular, Guttman's updating algorithm and Pliner's smoothing algorithm 
are discussed. A new method for finding a good starting value for these 
algorithms is proposed. The UDS solution provides an alternative method 
for cluster analysis. A method for automatic detection of cluster is
proposed . Finally, extensions and potential applications of this approach
to data mining will be discussed.