A maximum test in lieu of forcing a choice between the two dependent samples t-test and Wilcoxon signed-ranks test is proposed. The maximum test, which requires a new table of critical values, maintains nominal α whi...A maximum test in lieu of forcing a choice between the two dependent samples t-test and Wilcoxon signed-ranks test is proposed. The maximum test, which requires a new table of critical values, maintains nominal α while guaranteeing the maximum power of the two constituent tests. Critical values, obtained via Monte Carlo methods, are uniformly smaller than the Bonferroni-Dunn adjustment, giving it power superiority when testing for treatment alternatives of shift in location parameter when data are sampled from non-normal distributions.展开更多
We propose a new nonparametric test based on the rank difference between the paired sample for testing the equality of the marginal distributions from a bivariate distribution. We also consider a modification of the n...We propose a new nonparametric test based on the rank difference between the paired sample for testing the equality of the marginal distributions from a bivariate distribution. We also consider a modification of the novel nonparametric test based on the test proposed by Baumgartern, Weiβ, and Schindler (1998). An extensive numerical power comparison for various parametric and nonparametric tests was conducted under a wide range of bivariate distributions for small sample sizes. The two new nonparametric tests have comparable power to the paired t test for the data simulated from bivariate normal distributions, and are generally more powerful than the paired t test and other commonly used nonparametric tests in several important bivariate distributions.展开更多
The artificial bee colony (ABC) algorithm is a sim- ple and effective global optimization algorithm which has been successfully applied in practical optimization problems of various fields. However, the algorithm is...The artificial bee colony (ABC) algorithm is a sim- ple and effective global optimization algorithm which has been successfully applied in practical optimization problems of various fields. However, the algorithm is still insufficient in balancing ex- ploration and exploitation. To solve this problem, we put forward an improved algorithm with a comprehensive search mechanism. The search mechanism contains three main strategies. Firstly, the heuristic Gaussian search strategy composed of three different search equations is proposed for the employed bees, which fully utilizes and balances the exploration and exploitation of the three different search equations by introducing the selectivity probability P,. Secondly, in order to improve the search accuracy, we propose the Gbest-guided neighborhood search strategy for onlooker bees to improve the exploitation performance of ABC. Thirdly, the self- adaptive population perturbation strategy for the current colony is used by random perturbation or Gaussian perturbation to en- hance the diversity of the population. In addition, to improve the quality of the initial population, we introduce the chaotic opposition- based learning method for initialization. The experimental results and Wilcoxon signed ranks test based on 27 benchmark func- tions show that the proposed algorithm, especially for solving high dimensional and complex function optimization problems, has a higher convergence speed and search precision than ABC and three other current ABC-based algorithms.展开更多
With a random sample of 10 JCR(Science) subject areas it is shown that the 2-year and the 5-year impact factor of journals lead statistically to the same ranking per category. Yet in a majority of cases, the 5-year im...With a random sample of 10 JCR(Science) subject areas it is shown that the 2-year and the 5-year impact factor of journals lead statistically to the same ranking per category. Yet in a majority of cases, the 5-year impact factor is larger than the 2-year one.展开更多
小世界神经网络具有较快的收敛速度和优越的容错性,近年来得到广泛关注.然而,在网络构造过程中,随机重连可能造成重要信息丢失,进而导致网络精度下降.针对该问题,基于Watts-Strogatz(WS)型小世界神经网络,提出了一种基于突触巩固机制的...小世界神经网络具有较快的收敛速度和优越的容错性,近年来得到广泛关注.然而,在网络构造过程中,随机重连可能造成重要信息丢失,进而导致网络精度下降.针对该问题,基于Watts-Strogatz(WS)型小世界神经网络,提出了一种基于突触巩固机制的前馈小世界神经网络(Feedforward small-world neural network based on synaptic consolidation,FSWNN-SC).首先,使用网络正则化方法对规则前馈神经网络进行预训练,基于突触巩固机制,断开网络不重要的权值连接,保留重要的连接权值;其次,设计重连规则构造小世界神经网络,在保证网络小世界属性的同时实现网络稀疏化,并使用梯度下降算法训练网络;最后,通过4个UCI基准数据集和2个真实数据集进行模型性能测试,并使用Wilcoxon符号秩检验对对比模型进行显著性差异检验.实验结果表明:所提出的FSWNN-SC模型在获得紧凑的网络结构的同时,其精度显著优于规则前馈神经网络及其他WS型小世界神经网络.展开更多
文摘A maximum test in lieu of forcing a choice between the two dependent samples t-test and Wilcoxon signed-ranks test is proposed. The maximum test, which requires a new table of critical values, maintains nominal α while guaranteeing the maximum power of the two constituent tests. Critical values, obtained via Monte Carlo methods, are uniformly smaller than the Bonferroni-Dunn adjustment, giving it power superiority when testing for treatment alternatives of shift in location parameter when data are sampled from non-normal distributions.
文摘We propose a new nonparametric test based on the rank difference between the paired sample for testing the equality of the marginal distributions from a bivariate distribution. We also consider a modification of the novel nonparametric test based on the test proposed by Baumgartern, Weiβ, and Schindler (1998). An extensive numerical power comparison for various parametric and nonparametric tests was conducted under a wide range of bivariate distributions for small sample sizes. The two new nonparametric tests have comparable power to the paired t test for the data simulated from bivariate normal distributions, and are generally more powerful than the paired t test and other commonly used nonparametric tests in several important bivariate distributions.
基金supported by the Aviation Science Foundation of China(20105196016)the Postdoctoral Science Foundation of China(2012M521807)
文摘The artificial bee colony (ABC) algorithm is a sim- ple and effective global optimization algorithm which has been successfully applied in practical optimization problems of various fields. However, the algorithm is still insufficient in balancing ex- ploration and exploitation. To solve this problem, we put forward an improved algorithm with a comprehensive search mechanism. The search mechanism contains three main strategies. Firstly, the heuristic Gaussian search strategy composed of three different search equations is proposed for the employed bees, which fully utilizes and balances the exploration and exploitation of the three different search equations by introducing the selectivity probability P,. Secondly, in order to improve the search accuracy, we propose the Gbest-guided neighborhood search strategy for onlooker bees to improve the exploitation performance of ABC. Thirdly, the self- adaptive population perturbation strategy for the current colony is used by random perturbation or Gaussian perturbation to en- hance the diversity of the population. In addition, to improve the quality of the initial population, we introduce the chaotic opposition- based learning method for initialization. The experimental results and Wilcoxon signed ranks test based on 27 benchmark func- tions show that the proposed algorithm, especially for solving high dimensional and complex function optimization problems, has a higher convergence speed and search precision than ABC and three other current ABC-based algorithms.
基金supported by the National Natural Science Foundation of China(Grant No.70673019)
文摘With a random sample of 10 JCR(Science) subject areas it is shown that the 2-year and the 5-year impact factor of journals lead statistically to the same ranking per category. Yet in a majority of cases, the 5-year impact factor is larger than the 2-year one.
文摘小世界神经网络具有较快的收敛速度和优越的容错性,近年来得到广泛关注.然而,在网络构造过程中,随机重连可能造成重要信息丢失,进而导致网络精度下降.针对该问题,基于Watts-Strogatz(WS)型小世界神经网络,提出了一种基于突触巩固机制的前馈小世界神经网络(Feedforward small-world neural network based on synaptic consolidation,FSWNN-SC).首先,使用网络正则化方法对规则前馈神经网络进行预训练,基于突触巩固机制,断开网络不重要的权值连接,保留重要的连接权值;其次,设计重连规则构造小世界神经网络,在保证网络小世界属性的同时实现网络稀疏化,并使用梯度下降算法训练网络;最后,通过4个UCI基准数据集和2个真实数据集进行模型性能测试,并使用Wilcoxon符号秩检验对对比模型进行显著性差异检验.实验结果表明:所提出的FSWNN-SC模型在获得紧凑的网络结构的同时,其精度显著优于规则前馈神经网络及其他WS型小世界神经网络.