Joshua Huang (E-Business Technology Institute, HKU) 

Title: Clustering Large Data Sets in Data Mining and Cluster Validation

Abstract: Data mining is a process of discovering new knowledge from very 
large real world databases. Being scalable and being able to handle 
different data types are the two basic requirements for data mining 
algorithms. In this talk, I will focus clustering, one of the fundamental 
operations in data mining. I will first give a brief review of 
clustering algorithms used in data mining. Then, I will discuss the family 
of the k-means, k-modes and k-prototypes clustering algorithms. After 
that, I will discuss the problem of cluster validation in data mining and 
present a visual cluster validation method. Then, I will present a method to 
use the k-prototypes algorithm and visual cluster validation to interactively 
build classification models. Finally, I conclude my talk with some future 
research topics.