In non-smooth optimization,one particular problem which often appears inengineering designs,electrical engineering and game theory is called nonlinear minimaxproblem.For the non-smooth properties of objective function...In non-smooth optimization,one particular problem which often appears inengineering designs,electrical engineering and game theory is called nonlinear minimaxproblem.For the non-smooth properties of objective functions,there are some difficultiesin solving this problem.Since 1987,taking into account the entropy funtions,experts havehad several excellent results such as refs.[1—5].However,those methods are limited展开更多
孪生支持向量回归机(Twin Support Vector Regression,TSVR)的数学模型是求解一对约束优化问题,如何将约束优化问题转化为无约束优化问题进行求解是一个难题.在TSVR约束优化模型的基础上,依据最优化理论提出TSVR的无约束优化问题.然而,...孪生支持向量回归机(Twin Support Vector Regression,TSVR)的数学模型是求解一对约束优化问题,如何将约束优化问题转化为无约束优化问题进行求解是一个难题.在TSVR约束优化模型的基础上,依据最优化理论提出TSVR的无约束优化问题.然而,无约束优化问题的目标函数有可能不可微,为解决这个问题,引入极大熵函数,确保优化问题都是可微的.标准的极大熵函数法有可能发生数值溢出,所以对极大熵函数法进行了改进,提出自适应调节极大熵函数法来逼近TSVR的不可微项,并提出基于自适应调节极大熵函数法的TSVR学习算法.实验结果表明,和其他回归方法相比,所提算法不仅能够提高回归精度,而且效率得到了较大的提高.展开更多
基金Project supported by the National Natural Science Foundation of China.
文摘In non-smooth optimization,one particular problem which often appears inengineering designs,electrical engineering and game theory is called nonlinear minimaxproblem.For the non-smooth properties of objective functions,there are some difficultiesin solving this problem.Since 1987,taking into account the entropy funtions,experts havehad several excellent results such as refs.[1—5].However,those methods are limited
文摘孪生支持向量回归机(Twin Support Vector Regression,TSVR)的数学模型是求解一对约束优化问题,如何将约束优化问题转化为无约束优化问题进行求解是一个难题.在TSVR约束优化模型的基础上,依据最优化理论提出TSVR的无约束优化问题.然而,无约束优化问题的目标函数有可能不可微,为解决这个问题,引入极大熵函数,确保优化问题都是可微的.标准的极大熵函数法有可能发生数值溢出,所以对极大熵函数法进行了改进,提出自适应调节极大熵函数法来逼近TSVR的不可微项,并提出基于自适应调节极大熵函数法的TSVR学习算法.实验结果表明,和其他回归方法相比,所提算法不仅能够提高回归精度,而且效率得到了较大的提高.
基金The Major State Basic Research Development Program(2011CB302903)the China Postdoctoral Science Foundation(20100481167)+2 种基金the National Natural Science Foundation of China(60971129)the Natural Science Foundation of Jiangsu Province(BK2011793)the Postdoctoral Science Foundation of Jiangsu Province(1101022B)
基金Supported by the National High Technology Research and Development Program of China (863 Project)(2006AA010102)the Young Teacher Academic Fund of Nanjing University of Technology
