A Novel Approach
A Novel Approach
for Mixed Data Clustering using Dynamic Growing Hierarchical Self-Organizing
Map and Extended Attribute-Oriented Induction
Hari Prasad1 , M. Punithavalli2
1Assistant Professor(Selection Grade),
Department of Computer Applications, Sri Ramakrishna Institute of
Technology, Coimbatore, India.
2Director, Department of Computer
Applications, Sri Ramakrishna Engineering College, Coimbatore, India.
E-mail: hari.rese[email protected]
Abstract: Data clustering is one of the most important data
mining techniques which groups data supported on their similarity. A number of
approaches are existing for clustering numerical data and the problem of clustering
mixed data is still unresolved. The standard clustering techniques are in
general used for numeric data and are not probable to handle mixed data for the
reason that of their computational incompetence. The requisite for an enhanced
mixed data clustering approach is becoming vital and it is turning out to be a
hot research area. By the sort of resolving this issue, Growing Hierarchical
Self-Organizing Map (GHSOM) and Extended Attribute-Oriented Induction (EAOI)
for clustering mixed data type is previously projected except it does not have
any capability to control the growth of the map and in addition the structure
of GHSOM is static. To overcoming this issue, a Dynamic Growing Hierarchical
Self-Organizing Map (DGHSOM) with EAOI is projected in this paper for handling
the mixed data. The main importance of DGHSOM is that it has the ability to
grow or modify the structure to represent the application enhanced. The
experimentation for the proposed technique is approved with the help of UCI
Adult Data Set and Cleve Dataset and it is fond that it is superior to previous
approaches based on the number of resultant clusters and outliers with
substantial reduction in the processing time. The Clustering error also
reduced.
[Hari
Prasad, M. Punithavalli. A
Novel Approach for Mixed Data Clustering using Dynamic Growing Hierarchical
Self-Organizing Map and Extended Attribute-Oriented Induction. Life Sci J 2013:10(1):3259-3266]. (ISSN:
1097-8135). http://www.lifesciencesite.com.
Keywords: Mixed Data Clustering, Extended Attribute-Oriented Induction (EAOI), Self-Organizing Map, Dynamic Growing Hierarchical Self-Organizing Map (DGHSOM), Controlled Growth.
