Support vector machines (SVMs) are utilized for emotion recognition in Chinese speech in this paper. Both binary class discrimination and the multi class discrimination are discussed. It proves that the emotional fe...Support vector machines (SVMs) are utilized for emotion recognition in Chinese speech in this paper. Both binary class discrimination and the multi class discrimination are discussed. It proves that the emotional features construct a nonlinear problem in the input space, and SVMs based on nonlinear mapping can solve it more effectively than other linear methods. Multi class classification based on SVMs with a soft decision function is constructed to classify the four emotion situations. Compared with principal component analysis (PCA) method and modified PCA method, SVMs perform the best result in multi class discrimination by using nonlinear kernel mapping.展开更多
A SVMs (Support Vector Machines) based method to identify Chinese place names is presented. In our approach, place name candidate is located according to a rational forming assumption, then SVMs based identification s...A SVMs (Support Vector Machines) based method to identify Chinese place names is presented. In our approach, place name candidate is located according to a rational forming assumption, then SVMs based identification strategy is used to distinguish whether one candidate is true place name or not. Referring to linguistic knowledge, basic semanteme of a contextual word and frequency information of words inside place name candidate are selected as features in our methodology. So dimension in the feature space is reduced dramatically and processing procedure is performed more efficiently. Result of open testing on unregistered place names achieves F-measure 83.25 in 8.17 million words news based on this project.展开更多
Frequency-hopping(FH)is one of the commonly used spread spectrum techniques that finds wide applications in communications and radar systems because of its inherent capability of low interception,good confidentiality,...Frequency-hopping(FH)is one of the commonly used spread spectrum techniques that finds wide applications in communications and radar systems because of its inherent capability of low interception,good confidentiality,and strong antiinterference.However,non-cooperation FH transmitter classification is a significant and challenging issue for FH transmitter fingerprint feature recognition,since it not only is sensitive to noise but also has non-linear,non-Gaussian,and non-stability characteristics,which make it difficult to guarantee the classification in the original signal space.Some existing classifiers,such as the sparse representation classifier(SRC),generally use an individual representation rather than all the samples to classify the test data,which over-emphasizes sparsity but ignores the collaborative relationship among the given set of samples.To address these problems,we propose a novel classifier,called the kernel joint representation classifier(KJRC),for FH transmitter fingerprint feature recognition,by integrating kernel projection,collaborative feature representation,and classifier learning into a joint framework.Extensive experiments on real-world FH signals demonstrate the effectiveness of the proposed method in comparison with several state-of-the-art recognition methods.展开更多
We study tile local linear estimator for tile drift coefficient of stochastic differential equations driven by α-stable Levy motions observed at discrete instants. Under regular conditions, we derive the weak consis-...We study tile local linear estimator for tile drift coefficient of stochastic differential equations driven by α-stable Levy motions observed at discrete instants. Under regular conditions, we derive the weak consis- tency and central limit theorem of the estimator. Compared with Nadaraya-Watson estimator, the local linear estimator has a bias reduction whether the kernel function is symmetric or not under different schemes. A silnu- lation study demonstrates that the local linear estimator performs better than Nadaraya-Watson estimator, especially on the boundary.展开更多
文摘Support vector machines (SVMs) are utilized for emotion recognition in Chinese speech in this paper. Both binary class discrimination and the multi class discrimination are discussed. It proves that the emotional features construct a nonlinear problem in the input space, and SVMs based on nonlinear mapping can solve it more effectively than other linear methods. Multi class classification based on SVMs with a soft decision function is constructed to classify the four emotion situations. Compared with principal component analysis (PCA) method and modified PCA method, SVMs perform the best result in multi class discrimination by using nonlinear kernel mapping.
基金Foundation of China(Grant No.60175020and60673037) and the National High Technology Research and Development Program of China (Grant No.2002AA117010-09).
文摘A SVMs (Support Vector Machines) based method to identify Chinese place names is presented. In our approach, place name candidate is located according to a rational forming assumption, then SVMs based identification strategy is used to distinguish whether one candidate is true place name or not. Referring to linguistic knowledge, basic semanteme of a contextual word and frequency information of words inside place name candidate are selected as features in our methodology. So dimension in the feature space is reduced dramatically and processing procedure is performed more efficiently. Result of open testing on unregistered place names achieves F-measure 83.25 in 8.17 million words news based on this project.
基金Project supported by the National Natural Science Foundation of China(No.61601500)
文摘Frequency-hopping(FH)is one of the commonly used spread spectrum techniques that finds wide applications in communications and radar systems because of its inherent capability of low interception,good confidentiality,and strong antiinterference.However,non-cooperation FH transmitter classification is a significant and challenging issue for FH transmitter fingerprint feature recognition,since it not only is sensitive to noise but also has non-linear,non-Gaussian,and non-stability characteristics,which make it difficult to guarantee the classification in the original signal space.Some existing classifiers,such as the sparse representation classifier(SRC),generally use an individual representation rather than all the samples to classify the test data,which over-emphasizes sparsity but ignores the collaborative relationship among the given set of samples.To address these problems,we propose a novel classifier,called the kernel joint representation classifier(KJRC),for FH transmitter fingerprint feature recognition,by integrating kernel projection,collaborative feature representation,and classifier learning into a joint framework.Extensive experiments on real-world FH signals demonstrate the effectiveness of the proposed method in comparison with several state-of-the-art recognition methods.
基金supported by National Natural Science Foundation of China(Grant Nos.11171303 and 11071213)the Specialized Research Fund for the Doctor Program of Higher Education(Grant No.20090101110020)
文摘We study tile local linear estimator for tile drift coefficient of stochastic differential equations driven by α-stable Levy motions observed at discrete instants. Under regular conditions, we derive the weak consis- tency and central limit theorem of the estimator. Compared with Nadaraya-Watson estimator, the local linear estimator has a bias reduction whether the kernel function is symmetric or not under different schemes. A silnu- lation study demonstrates that the local linear estimator performs better than Nadaraya-Watson estimator, especially on the boundary.
