In order to overcome the high dimension and collinearity problem of spectrum data in spectroscopy quantitative analysis, we introduce a partial least squares support vector machines (PLS-SVM) method, which integrates ...
详细信息
In order to overcome the high dimension and collinearity problem of spectrum data in spectroscopy quantitative analysis, we introduce a partial least squares support vector machines (PLS-SVM) method, which integrates partial least squares (PLS) and support vector machines (SVM). The method uses PLS to extract the feature of spectrum. Then the feature serves as the input of SVM calibration model instead of the whole spectrum. Experimental results show that PLS-SVM is superior to PLS in terms of prediction precision, while the modeling time is greatly reduced compared with SVM without feature extraction
This paper proposes a development method to deal with the stability problems for a kind of complex large-scale systems with hybrid models. These hybrid large-scale systems are composed of considerable interconnected n...
详细信息
This paper proposes a development method to deal with the stability problems for a kind of complex large-scale systems with hybrid models. These hybrid large-scale systems are composed of considerable interconnected nonlinear subsystems, some of which are described by differential equations and others by Takagi-Sugeno fuzzy models. Novel techniques to cope with the nonlinear interconnection between subsystems are developed. A set of LMI-based conditions are derived to judge the stability of the whole system by checking the stability of the subsystems in parallel, which greatly speeds up the stability analysis process. And the computational complexity is greatly reduced. A numerical example is given to illustrate its effectiveness
In the standard support vector machines for classification, training sets with uneven class sizes results in classification biases towards the class with the large training size. That is to say, the larger the trainin...
详细信息
In the standard support vector machines for classification, training sets with uneven class sizes results in classification biases towards the class with the large training size. That is to say, the larger the training sample size for one class is, the smaller its corresponding classification error rate is, while the smaller the sample size, the larger the classification error rate. The main causes lie in that the penalty of misclassification for each training sample is considered equally. Weighted support vector machines for classification are proposed in this paper where penalty of misclassification for each training sample is different. By setting the equal penalty for the training samples belonging to same class, and setting the ratio of penalties for different classes to the inverse ratio of the training class sizes, the obtained weighted support vector machines compensate for the undesirable effects caused by the uneven training class size, and the classification accuracy for the class with small training size is improved. Experimental simulations on breast cancer diagnosis show the effectiveness of the proposed methods.
In the standard support vector machines for classification, the use of training sets with uneven class sizes results in classification biases towards the class with the large training size. The main causes lie in that...
详细信息
In the standard support vector machines for classification, the use of training sets with uneven class sizes results in classification biases towards the class with the large training size. The main causes lie in that the penalty of misclassification for each training sample is considered equally. Weighted support vector machines for classification are proposed in this paper where penalty of misclassification for each training sample is different. By setting the equal penalty for the training samples belonging to same class, and setting the ratio of penalties for different classes to the inverse ratio of the training class sizes, the obtained weighted support vector machines compensate for the undesirable effects caused by the uneven training class size, and the classification accuracy for the class with small training size is improved. But this improvement is obtained at the cost of the possible decrease of classification accuracy for the class with large training size and the possible decrease of the total classification accuracy. Two weighted support vector machines, namely weighted C-SVM and V-SVM, corresponding to C-SVM and V-SVM are given respectively. Experimental simulations on breast cancer diagnosis show the effectiveness of the proposed methods.
A delay-dependent sufficient condition for the existence of a robust H/sub /spl infin// switched controller with state delay feedback for linear switched systems with parameter uncertainties and time delay was derived...
详细信息
ISBN:
(纸本)0780382730
A delay-dependent sufficient condition for the existence of a robust H/sub /spl infin// switched controller with state delay feedback for linear switched systems with parameter uncertainties and time delay was derived and formulated in nonlinear matrix inequalities solvable by an iterative algorithm. Compared with the conventional memoryless state-feedback controller, the proposed controller can achieve better robust control performance since the delayed state is utilized as additional feedback information and the parameters of the proposed controllers are changed synchronously with the dynamical characteristic of the system. This design method was also extended to the case where only delayed state is available for the controller. The example of balancing an inverted pendulum on a cart demonstrates the effectiveness and applicability of the proposed design methods.
暂无评论