The Experts below are selected from a list of 97893 Experts worldwide ranked by ideXlab platform
Deqing Wang - One of the best experts on this subject based on the ideXlab platform.
-
unsupervised feature selection through gram schmidt orthogonalization a word co occurrence perspective
Neurocomputing, 2016Co-Authors: Deqing Wang, Hui Zhang, Jing WangAbstract:Feature selection is a key step in many machine learning applications, such as categorization, and clustering. Especially for text data, the original document-Term Matrix is high-dimensional and sparse, which affects the performance of feature selection algorithms. Meanwhile, labeling training instance is time-consuming and expensive. So unsupervised feature selection algorithms have attracted more attention. In this paper, we propose an unsupervised feature selection algorithm through R ? andom P ? rojection and G ? ram- G ? chmidt O ? rthogonalization (RP-GSO) from the word co-occurrence Matrix. The RP-GSO algorithm has three advantages: (1) it takes as input dense word co-occurrence Matrix, avoiding the sparseness of original document-Term Matrix; (2) it selects "basis features" by Gram-Schmidt process, guaranteeing the orthogonalization of feature space; and (3) it adopts random projection to speed up GS process. Extensive experimental results show our proposed RP-GSO approach achieves better performance comparing against supervised and unsupervised feature selection methods in text classification and clustering tasks.
-
ICDM - GS-Orthogonalization Based "Basis Feature" Selection from Word Co-occurrence Matrix
2015 IEEE International Conference on Data Mining, 2015Co-Authors: Deqing Wang, Hui Zhang, Rui LiuAbstract:Feature selection plays an important role in machinelearning applications. Especially for text data, the highdimensionaland sparse characteristics will affect the performanceof feature selction. In this paper, an unsupervised feature selection algorithm through Random Projection and Gram-Schmidt Orthogonalization (RP-GSO) from the word co-occurrence Matrix is proposed. The RP-GSO has three advantages: (1) it takes as input dense word co-occurrence Matrix, avoiding the sparseness of original document-Term Matrix, (2) it selects "basis features" by Gram-Schmidt process, guaranteeing the orthogonalization of feature space, and (3) it adopts random projection to speed upGS process. We did extensive experiments on two real-world textcorpora, and observed that RP-GSO achieves better performancecomparing against supervised and unsupervised methods in textclassification and clustering tasks.
Jing Wang - One of the best experts on this subject based on the ideXlab platform.
-
unsupervised feature selection through gram schmidt orthogonalization a word co occurrence perspective
Neurocomputing, 2016Co-Authors: Deqing Wang, Hui Zhang, Jing WangAbstract:Feature selection is a key step in many machine learning applications, such as categorization, and clustering. Especially for text data, the original document-Term Matrix is high-dimensional and sparse, which affects the performance of feature selection algorithms. Meanwhile, labeling training instance is time-consuming and expensive. So unsupervised feature selection algorithms have attracted more attention. In this paper, we propose an unsupervised feature selection algorithm through R ? andom P ? rojection and G ? ram- G ? chmidt O ? rthogonalization (RP-GSO) from the word co-occurrence Matrix. The RP-GSO algorithm has three advantages: (1) it takes as input dense word co-occurrence Matrix, avoiding the sparseness of original document-Term Matrix; (2) it selects "basis features" by Gram-Schmidt process, guaranteeing the orthogonalization of feature space; and (3) it adopts random projection to speed up GS process. Extensive experimental results show our proposed RP-GSO approach achieves better performance comparing against supervised and unsupervised feature selection methods in text classification and clustering tasks.
Hui Zhang - One of the best experts on this subject based on the ideXlab platform.
-
unsupervised feature selection through gram schmidt orthogonalization a word co occurrence perspective
Neurocomputing, 2016Co-Authors: Deqing Wang, Hui Zhang, Jing WangAbstract:Feature selection is a key step in many machine learning applications, such as categorization, and clustering. Especially for text data, the original document-Term Matrix is high-dimensional and sparse, which affects the performance of feature selection algorithms. Meanwhile, labeling training instance is time-consuming and expensive. So unsupervised feature selection algorithms have attracted more attention. In this paper, we propose an unsupervised feature selection algorithm through R ? andom P ? rojection and G ? ram- G ? chmidt O ? rthogonalization (RP-GSO) from the word co-occurrence Matrix. The RP-GSO algorithm has three advantages: (1) it takes as input dense word co-occurrence Matrix, avoiding the sparseness of original document-Term Matrix; (2) it selects "basis features" by Gram-Schmidt process, guaranteeing the orthogonalization of feature space; and (3) it adopts random projection to speed up GS process. Extensive experimental results show our proposed RP-GSO approach achieves better performance comparing against supervised and unsupervised feature selection methods in text classification and clustering tasks.
-
ICDM - GS-Orthogonalization Based "Basis Feature" Selection from Word Co-occurrence Matrix
2015 IEEE International Conference on Data Mining, 2015Co-Authors: Deqing Wang, Hui Zhang, Rui LiuAbstract:Feature selection plays an important role in machinelearning applications. Especially for text data, the highdimensionaland sparse characteristics will affect the performanceof feature selction. In this paper, an unsupervised feature selection algorithm through Random Projection and Gram-Schmidt Orthogonalization (RP-GSO) from the word co-occurrence Matrix is proposed. The RP-GSO has three advantages: (1) it takes as input dense word co-occurrence Matrix, avoiding the sparseness of original document-Term Matrix, (2) it selects "basis features" by Gram-Schmidt process, guaranteeing the orthogonalization of feature space, and (3) it adopts random projection to speed upGS process. We did extensive experiments on two real-world textcorpora, and observed that RP-GSO achieves better performancecomparing against supervised and unsupervised methods in textclassification and clustering tasks.
Kai-yuan Cai - One of the best experts on this subject based on the ideXlab platform.
-
Stability Analysis of a Class of Neutral Type Systems in a Critical Case Without Restriction on the Principal Neutral Term
Asian Journal of Control, 2012Co-Authors: Quan Quan, Kai-yuan CaiAbstract:This study focuses mainly on the stability of a class of linear neutral systems in a critical case, that is, where the spectral radius of the principal neutral Term (Matrix H in this paper) is equal to 1. It is difficult to deTermine the stability of such systems via existing methods. In this study, a sufficient stability criterion for the critical case without restrictions on the principal neutral Term is given in Terms of the existence of solutions to a linear Matrix inequality.
-
A stability theorem of the direct Lyapunov's method for neutral-type systems in a critical case
International Journal of Systems Science, 2012Co-Authors: Quan Quan, Kai-yuan CaiAbstract:A new stability theorem of the direct Lyapunov's method is proposed for neutral-type systems. The main contribution of the proposed theorem is to remove the condition that the 𝒟 operator is stable. In order to demonstrate the effectiveness, the proposed theorem is used to deTermine the stability of a neutral-type system in a critical case, i.e. the dominant eigenvalues of the principal neutral Term (Matrix D in Introduction) lie on the unit circle. This is difficult or infeasible in previous studies.
-
Robustness analysis for a class of linear neutral systems in a critical case
IET Control Theory & Applications, 2010Co-Authors: Quan Quan, Kai-yuan CaiAbstract:This study pays attention to robust stability of a class of uncertain neutral systems in a critical case, where the spectral radius of the principal neutral Term (Matrix H in this study) is equal to 1. It is shown that usual methods cannot deal with such systems with uncertainties. Thus, a novel method is developed, whose idea is to examine whether or not an existing stability criterion still holds when the uncertainties are sufficiently small. More specifically, it is to examine whether or not the existing stability criterion in Terms of a linear Matrix inequality (LMI) still has a solution with the sufficiently small uncertainties. By analysing the structure of the solution, a new robust stability criterion is derived in Terms of the existence of solutions to an equation. An application shows the effectiveness of the proposed method by dealing with a problem caused by numerical methods when solving LMIs.
-
A framework for the stability of a class of neutral type systems in a critical case
2010Co-Authors: Quan Quan, Kai-yuan CaiAbstract:Under the proposed framework, a class of existing stability criteria can be extended to deTermine the stability of a class of neutral type systems in a critical case, i.e., the spectral radius of the principal neutral Term (Matrix D in this paper) is equal to 1. Furthermore, an interesting property of the extended criteria is given. Illustrative examples are presented to demonstrate the effectiveness of the proposed framework.
-
Brief paper: Linear Matrix inequality approach for stability analysis of linear neutral systems in a critical case
IET Control Theory & Applications, 2010Co-Authors: Quan Quan, De-dong Yang, Kai-yuan CaiAbstract:This study mainly focuses on the stability of a class of linear neutral systems in a critical case, that is, the spectral radius of the principal neutral Term (Matrix H) is equal to 1. It is difficult to deTermine the stability of such systems by using existing methods. In this study, a sufficient stability criterion for the critical case is given in Terms of the existence of solutions to a linear Matrix inequality (LMI). Moreover, it is also shown that the proposed stability criterion conforms with a fact that the considered linear neutral systems are unstable when H has a Jordan block corresponding to the eigenvalue of modulus 1. An illustrative example is presented to deTermine the stability of a linear neutral system whose principal neutral Term H has multiple eigenvalues of modulus 1 without Jordan chains. This is difficult in existing studies.
Wangjing - One of the best experts on this subject based on the ideXlab platform.
-
Unsupervised feature selection through Gram-Schmidt orthogonalization-A word co-occurrence perspective
Neurocomputing, 2016Co-Authors: Wangdeqing, Zhanghui, Liurui, Liuxianglong, WangjingAbstract:Feature selection is a key step in many machine learning applications, such as categorization, and clustering. Especially for text data, the original document-Term Matrix is high-dimensional and sp...