{
"title": "The Robust Clustering with Reduction Dimension",
"authors": "Dyah E. Herwindiati",
"country": null,
"institution": null,
"volume": "63",
"journal": "International Journal of Mathematical, Computational, Physical, Electrical and Computer Engineering",
"pagesStart": 199,
"pagesEnd": 205,
"ISSN": "1307-6892",
"URL": "http:\/\/waset.org\/publications\/14058",
"abstract": "A clustering is process to identify a homogeneous\ngroups of object called as cluster. Clustering is one interesting topic\non data mining. A group or class behaves similarly characteristics.\nThis paper discusses a robust clustering process for data images with\ntwo reduction dimension approaches; i.e. the two dimensional\nprincipal component analysis (2DPCA) and principal component\nanalysis (PCA). A standard approach to overcome this problem is\ndimension reduction, which transforms a high-dimensional data into\na lower-dimensional space with limited loss of information. One of\nthe most common forms of dimensionality reduction is the principal\ncomponents analysis (PCA). The 2DPCA is often called a variant of\nprincipal component (PCA), the image matrices were directly treated\nas 2D matrices; they do not need to be transformed into a vector so\nthat the covariance matrix of image can be constructed directly using\nthe original image matrices. The decomposed classical covariance\nmatrix is very sensitive to outlying observations. The objective of\npaper is to compare the performance of robust minimizing vector\nvariance (MVV) in the two dimensional projection PCA (2DPCA)\nand the PCA for clustering on an arbitrary data image when outliers\nare hiden in the data set. The simulation aspects of robustness and\nthe illustration of clustering images are discussed in the end of\npaper",
"references": null,
"publisher": "World Academy of Science, Engineering and Technology",
"index": "International Science Index 63, 2012"
}