Efficient and robust feature extraction by maximum margin criterion.

Efficient and robust feature extraction by maximum margin criterion. Research Abstract Details 

Research Abstract Table of Contents

Jump to the:

Efficient and robust feature extraction by maximum margin criterion. Abstract Text:

Jun Liu,Songcan Chen,Xiaoyang Tan,Daoqiang Zhang,

In pattern recognition, feature extraction techniques are widely employed to reduce the dimensionality of data and to enhance the discriminatory information. Principal component analysis (PCA) and linear discriminant analysis (LDA) are the two most popular linear dimensionality reduction methods. However, PCA is not very effective for the extraction of the most discriminant features, and LDA is not stable due to the small sample size problem. In this paper, we propose some new (linear and nonlinear) feature extractors based on maximum margin criterion (MMC). Geometrically, feature extractors based on MMC maximize the (average) margin between classes after dimensionality reduction. It is shown that MMC can represent class separability better than PCA. As a connection to LDA, we may also derive LDA from MMC by incorporating some constraints. By using some other constraints, we establish a new linear feature extractor that does not suffer from the small sample size problem, which is known to cause serious stability problems for LDA. The kernelized (nonlinear) counterpart of this linear feature extractor is also established in the paper. Our extensive experiments demonstrate that the new feature extractors are effective, stable, and efficient.

Efficient and robust feature extraction by maximum margin criterion. Publishing Authors By Initials

J Liu,S Chen,X Tan,D Zhang,

For similar information science: pattern recognition, automated research abstracts see: information science: pattern recognition, automated research

PUBMED ID PMID:

MEDLINE DATE:

Efficient and robust feature extraction by maximum margin criterion. Journal Published:

PUBLICATION TYPE: Research Support, U.S. Gov't,

Journal: IEEE transactions on neural networks / a publicati

VOLUME: 17

Page Numbers: 157-65

Journal Abbreviation:

ISSN: 1045-9227

DAY: 4

MONTH: Jan

YEAR: 2006

Efficient and robust feature extraction by maximum margin criterion. Information

Number of References:

LANGUAGE: eng

NlmUniqueID: 101211035

Efficient and robust feature extraction by maximum margin criterion. Keywords Mesh Terms:

KEYWORDS: Pattern Recognition, Automated

MESH TERMS:

Chemical & Substance for Abstract: Efficient and robust feature extraction by maximum margin criterion. Information

Substance Name:

Registry Number:

Grant and Affiliation Information for Efficient and robust feature extraction by maximum margin criterion.

AFFILIATION: Department of Computer Science and Engineering, University of California, Riverside, CA 92521, USA. hli@cs.ucr.edu

Country: United States

AGENCY:

GRANT:

ACRONYM:

MEDLINETA: IEEE Trans Neural Netw

REFSOURCE: IEEE Trans Neural Netw. 2007 Nov;18(6):1

DATABASENAME:

ACCESSION NUMBER:

Number Hits: 0

Efficient and robust feature extraction by maximum margin criterion Related Publications

• A comparative study on the intercuspal occlusion among TMD patients, malocclusion patients and university students.
• Effect of retinoblastoma gene on human osteosarcoma cell line OS732.
• Enhanced antioxidant and cytoprotective abilities of vitamin E succinate is associated with a rapid uptake advantage in rat hepatocytes and mitochondria.
• Intervention by azithromycin on bleomycin-induced lung injury in rats and its mechanisms.
• Method development of adsorbed stationary phase open tubular capillary electrochromatography.
• Supercritical fluid CO2 extraction of colchicine from colchicine Autumnele liliaceae.
• Vascular endothelial growth factor gene expression in patients' ischemic skeletal muscle with diabetic foot.
• [A case of hypodermiasis with empyema]
• [Analysis of hepatitis G virus genotypes in some areas of China]
• [Autologous dendritic cells eliciting cytotoxicity of bone marrow cells against chronic granulocytic leukemia]
• [Changes of human recombination bone morphogenetic protein-2 in bone and marrow in tail suspended rats]
• [Cloning of the mouse Doc-1R gene by genomic walking]
• [Construction of recombinant adenoviruses carrying human urokinase-type plasminogen activator and its expression in hepatic stellate cells in vitro]
• [Effect of inhibition of telomerase activity on cisplatin-induced apoptosis in K562 cells]
• [Effects of electrical stimulation of lateral hypothalamic area on gastric ischemia-reperfusion injury in rats]
• [Fatal complications in traffic accident injury]
• [Indomethacin induces apoptosis of K562 cells through activation of caspases and elevation of intracellular free calcium]
• [Infection and replication of hepatitis B virus and its clearance mechanism: an overview]
• [Mechanisms underlying dopaminergic neurotoxicity in ischemic cerebral damage]
• [Physiological and pathologic implications for zero-stress state of the esophagus]
• [Protein kinases and Alzheimer's disease]
• [Retroperitoneal laparoscopic unroofing of renal cyst: use of close method]
• [Risk of leukemia among medical X-ray workers in China between 1950 and 1995]
• [Studies on association between lipoprotein lipase gene polymorphisms of Pvu II site and hypertriglyceridemics in Chinese]
• [Study on automatic segmentation of color images applied to blood cells]
• [Study on high performance liquid chromatography/indirect photometric detection of fosfomycin with acridine as detection reagent]
• [Study on the application of number fluorescence density in detecting autoantibodies titer]
• [The effects of surface morphology of calcium phosphate ceramics on apatite formation in dynamic SBF]
• [Transforming growth factor beta(1) and bone morphogenetic protein 2 induce the differentiation of odontoblasts in vitro]
• [Vector propagation algorithm based ECG simulation of bundle branch block]