The Experts below are selected from a list of 324 Experts worldwide ranked by ideXlab platform
Feilong Cao - One of the best experts on this subject based on the ideXlab platform.
-
optimization Approximation Solution for regression problem based on extreme learning machine
Neurocomputing, 2011Co-Authors: Yubo Yuan, Yu Guang Wang, Feilong CaoAbstract:Abstract Extreme learning machine (ELM) is one of the most popular and important learning algorithms. It comes from single-hidden-layer feedforward neural networks. It has been proved that ELM can achieve better performance than support vector machine (SVM) in regression and classification. In this paper, mathematically, with regression problem, the step 3 of ELM is studied. First of all, the equation H β = T are reformulated as an optimal model. With the optimality, the necessary conditions of optimal Solution are presented. The equation H β = T is replaced by H T H β = H T T . We can prove that the latter must have one Solution at least. Second, optimal Approximation Solution is discussed in cases of H is column full rank, row full rank, neither column nor row full rank. In the last case, the rank-1 and rank-2 methods are used to get optimal Approximation Solution. In theory, this paper present a better algorithm for ELM.
-
sixth order polynomial smoothing Approximation Solution to support vector machine
International Conference on Machine Learning and Cybernetics, 2010Co-Authors: Yubo Yuan, Feilong CaoAbstract:Support vector machine (SVM) can be seen as a special binary classification method. The original model is a quadratical programming with linear inequalities constraints. It is a very important issue that how to get the optimal Solution of SVM model. In this paper, a new Solution method is proposed. The constraints are moved away from the original optimization model by using the Approximation Solution in the feasible space. Three points under one control parameter smoothing function is used to smoothen the objective function of unconstrained model. It is a sixth order polynomial function. The smoothing performance is investigated. By theory proof, the proposed unconstrained model has an active performance which can be controlled by one proposed parameter.
-
ICMLC - Sixth order polynomial smoothing Approximation Solution to support vector machine
2010 International Conference on Machine Learning and Cybernetics, 2010Co-Authors: Yubo Yuan, Feilong CaoAbstract:Support vector machine (SVM) can be seen as a special binary classification method. The original model is a quadratical programming with linear inequalities constraints. It is a very important issue that how to get the optimal Solution of SVM model. In this paper, a new Solution method is proposed. The constraints are moved away from the original optimization model by using the Approximation Solution in the feasible space. Three points under one control parameter smoothing function is used to smoothen the objective function of unconstrained model. It is a sixth order polynomial function. The smoothing performance is investigated. By theory proof, the proposed unconstrained model has an active performance which can be controlled by one proposed parameter.
Yubo Yuan - One of the best experts on this subject based on the ideXlab platform.
-
optimization Approximation Solution for regression problem based on extreme learning machine
Neurocomputing, 2011Co-Authors: Yubo Yuan, Yu Guang Wang, Feilong CaoAbstract:Abstract Extreme learning machine (ELM) is one of the most popular and important learning algorithms. It comes from single-hidden-layer feedforward neural networks. It has been proved that ELM can achieve better performance than support vector machine (SVM) in regression and classification. In this paper, mathematically, with regression problem, the step 3 of ELM is studied. First of all, the equation H β = T are reformulated as an optimal model. With the optimality, the necessary conditions of optimal Solution are presented. The equation H β = T is replaced by H T H β = H T T . We can prove that the latter must have one Solution at least. Second, optimal Approximation Solution is discussed in cases of H is column full rank, row full rank, neither column nor row full rank. In the last case, the rank-1 and rank-2 methods are used to get optimal Approximation Solution. In theory, this paper present a better algorithm for ELM.
-
sixth order polynomial smoothing Approximation Solution to support vector machine
International Conference on Machine Learning and Cybernetics, 2010Co-Authors: Yubo Yuan, Feilong CaoAbstract:Support vector machine (SVM) can be seen as a special binary classification method. The original model is a quadratical programming with linear inequalities constraints. It is a very important issue that how to get the optimal Solution of SVM model. In this paper, a new Solution method is proposed. The constraints are moved away from the original optimization model by using the Approximation Solution in the feasible space. Three points under one control parameter smoothing function is used to smoothen the objective function of unconstrained model. It is a sixth order polynomial function. The smoothing performance is investigated. By theory proof, the proposed unconstrained model has an active performance which can be controlled by one proposed parameter.
-
ICMLC - Sixth order polynomial smoothing Approximation Solution to support vector machine
2010 International Conference on Machine Learning and Cybernetics, 2010Co-Authors: Yubo Yuan, Feilong CaoAbstract:Support vector machine (SVM) can be seen as a special binary classification method. The original model is a quadratical programming with linear inequalities constraints. It is a very important issue that how to get the optimal Solution of SVM model. In this paper, a new Solution method is proposed. The constraints are moved away from the original optimization model by using the Approximation Solution in the feasible space. Three points under one control parameter smoothing function is used to smoothen the objective function of unconstrained model. It is a sixth order polynomial function. The smoothing performance is investigated. By theory proof, the proposed unconstrained model has an active performance which can be controlled by one proposed parameter.
J.l. Sedwick - One of the best experts on this subject based on the ideXlab platform.
-
Successive Approximation Solution of the HJI equation
Proceedings of 1994 33rd IEEE Conference on Decision and Control, 1994Co-Authors: K.a. Wise, J.l. SedwickAbstract:A successive Approximation Solution approach is used to solve the Hamilton-Jacobi-Isaacs partial differential Riccati equation that arises in nonlinear H/sub /spl infin// optimal control problems. The method of characteristics is used to form integral expressions used to solve the partial differential Riccati equation. Successive Approximations are then used to approximate the Solution to these integral expressions. Two examples in which the optimal control is known are presented. These examples outline the approximate Solution approach and provide insight into what degree of Approximation is needed.
Zhang Chengke - One of the best experts on this subject based on the ideXlab platform.
-
Analysis Method for Nash Strategy of Linear Time Variant Quadratic Differential Game via Wavelets (II) —Convergence of the Wavelet Approximation Solution
Control theory & applications, 2002Co-Authors: Zhang ChengkeAbstract:This paper studies the convergence problem of the wavelet Approximation analysis method. For Nash strategy of linear time variant quadratic differential game, we prove that the wavelet Approximation Solution of Nash strategy converge to the accurate Solution. The order of error estimation is given based on the multi_scale multi_reSolution Approximation feature of wavelets.
K.a. Wise - One of the best experts on this subject based on the ideXlab platform.
-
Successive Approximation Solution of the HJI equation
Proceedings of 1994 33rd IEEE Conference on Decision and Control, 1994Co-Authors: K.a. Wise, J.l. SedwickAbstract:A successive Approximation Solution approach is used to solve the Hamilton-Jacobi-Isaacs partial differential Riccati equation that arises in nonlinear H/sub /spl infin// optimal control problems. The method of characteristics is used to form integral expressions used to solve the partial differential Riccati equation. Successive Approximations are then used to approximate the Solution to these integral expressions. Two examples in which the optimal control is known are presented. These examples outline the approximate Solution approach and provide insight into what degree of Approximation is needed.