"Clustering algorithms for microarray data: overview and 
comparative studies." 
Susmita Datta, 
Department of Mathematics and  Statistics,  
Georgia State University

Abstract
Microarray chip has revolutionized the area of genomics. In 
the beginning, clustering was used to group genes according 
to the similarity of their expression profiles.  
Oftentimes, members in the same group had similar 
biological functions. The current emphasis of clustering 
seems to be in grouping tissue samples from the expression 
values. For example if microarray data are collected on 
different types of cancer cells then clustering those cells 
will provide information about the similarity of their gene 
expression profiles. This objective has the additional 
problem of dimensionality reduction of the underlying 
multivariate data. Given any of these clustering problems 
there are numerous clustering techniques to choose from. In 
this talk we will present an overview of some of these  
clustering algorithms. Their performances are examined on 
two well-known datasets for each of the two types of 
clustering problems. We introduce several validation 
measures to evaluate the performance of a clustering  
algorithm for a particular situation.