Mahidol University Logo
Faculty of ICT, Mahidol University
 

Admissions

Printable Version

 

K-MEANS CLUSTERING USING CORRELATED ATTRIBUTES

 

TITLE K-MEANS CLUSTERING USING CORRELATED ATTRIBUTES
AUTHOR THIPPRAPA POPIYATRAKUL
DEGREE MASTER OF SCIENCE PROGRAMME IN COMPUTER SCIENCE
FACULTY FACULTY OF SCIENCE
ADVISOR SUDSANGUAN NGAMSURIYAROJ
CO-ADVISOR SONGSRI TANGSRIPAIROJ
ANANTA SRISUPHAB
 

ABSTRACT

K-Means clustering is one of the widely used knowledge discovery techniques. One disadvantage of K-means clustering used for a large data set is how to find an initial set of clustered data that are approximately close to the final set of clustered data so that it would not take a lot of time in clustering the final set when compared with the method that selects a data set in a random fashion.

This thesis proposed a new efficient way to do the K-means clustering when only a few correlated attributes are used. For each data set, the correlation among attributes is computed in order to determine which attributes are most related so that they could be the representatives of all attributes. Subsequently, the correlated attributes are selected for computing the clustering, and the results are compared with the clustering outputs obtained from using all attributes in the computation.

We evaluated our work using 10 datasets of UCI Machine Learning Repository including Iris, Ecoli, Yeat, Abalone, White Wine, Page Blocks, Magic Grammar Telescope, Breast Cancer, Waveform, and Letter Recognition. We used the tool called PML (Parallel Machine Learning) developed by IBM for performing the K-means clustering. Our results illustrate that the centroids of correlated attributes are close to the centroids of all attributes while having less computation time.


KEYWORD K-MEANS CLUSTERING / INITIAL CENTROID / CORRELATION

 

 

Go to Top

 

ICT Building, Mahidol University, 999 Phuttamonthon 4 Road, Salaya, Nakhonpathom 73170 Tel. +66 02 441-0909 Fax. +66 02 849-6099
Mahidol University Computing Center, The Faculty of ICT, Mahidol University , Rama 6 Road, Rajathevi, Bangkok 10400 Tel. +66 02 354-4333 Fax. +66 02 354-7333