Study of Multivariate Data Clustering Based on K-Means and Independent Component Analysis
Open Access
- 1 January 2015
- journal article
- Published by Science Publishing Group in American Journal of Theoretical and Applied Statistics
- Vol. 4 (5) , 317-321
Abstract
For last two decades, clustering is well-recognized area in the research field of data mining. Data clustering plays the major research at pattern recognition, Signal processing, bioinformatics and Artificial Intelligence. Clustering process is an unsupervised learning techniques where it generates a group of object based on their similarity in such a way that the objects belonging to other groups are similar and those belonging to other are dissimilar. This paper analysis the three different data types clustering techniques like K-Means, Principal components analysis (PCA) and Independent component analysis (ICA) in real and simulated data. The recent developments by considering a rather unexpected application of the theory of Independent component analysis (ICA) found in data clustering, outlier detection and multivariate data visualization. Accurate identification of data clustering plays an important role in statistical analysis. In this paper we explore the connection among these three techniques to identify multivariate data clustering and develop a new method k-means on PCA or ICA then the result shows that ICA based clustering performs well than others.Keywords
This publication has 11 references indexed in Scilit:
- Multivariate Outlier Detection Using Independent Component AnalysisScience Journal of Applied Mathematics and Statistics, 2015
- Improving the performance of kurtosis estimatorComputational Statistics & Data Analysis, 2008
- Metabolite fingerprinting: detecting biological features by independent component analysisBioinformatics, 2004
- The Elements of Statistical Learning2001
- Independent component analysis: algorithms and applicationsNeural Networks, 2000
- A Class of Quantile Measures for KurtosisThe American Statistician, 1998
- Least squares quantization in PCMIEEE Transactions on Information Theory, 1982
- Algorithm AS 136: A K-Means Clustering AlgorithmJournal of the Royal Statistical Society Series C: Applied Statistics, 1979
- The Approximation of One Matrix by Another of Lower RankPsychometrika, 1936
- "Das Fehlergesetz und Seine Verallgemeinerungen Durch Fechner und Pearson." A RejoinderBiometrika, 1905