In this paper, we study the flocking behavior of a thermodynamic Cucker–Smale model with local velocity interactions. Using the spectral gap of a connected stochastic matrix, together with an elaborate estimate on pe...In this paper, we study the flocking behavior of a thermodynamic Cucker–Smale model with local velocity interactions. Using the spectral gap of a connected stochastic matrix, together with an elaborate estimate on perturbations of a linearized system, we provide a sufficient framework in terms of initial data and model parameters to guarantee flocking. Moreover, it is shown that the system achieves a consensus at an exponential rate.展开更多
智能电表故障的准确预测对实现计量设备精准主动运维、保障电网稳定运行具有重要意义。电表各故障类型样本的出现频次不同,且不同故障类型样本在高维特征空间中的分布存在重叠,这极大增加了故障预测的难度。现有不平衡分类方法通过构建...智能电表故障的准确预测对实现计量设备精准主动运维、保障电网稳定运行具有重要意义。电表各故障类型样本的出现频次不同,且不同故障类型样本在高维特征空间中的分布存在重叠,这极大增加了故障预测的难度。现有不平衡分类方法通过构建单一样本信息与其对应类别标签的映射关系来划分样本类型,导致对具有相似表征信息的重叠区样本难以准确判别,降低了整体分类精度。该文提出一种基于多粒度近邻图的智能电表故障分类方法。首先,选择原始数据集中样本作为目标样本,以目标样本及其近邻样本作为节点、目标样本与其近邻样本连线作为边构建近邻图。根据选择的近邻样本数量不同构建多粒度近邻图,实现目标样本的信息扩充和训练样本的数量扩增,更有利于模型稳定训练。构建编码器挖掘近邻图节点特征,利用图注意力机制,根据近邻图节点编码特征和节点邻接关系将近邻样本信息自适应地聚合到目标样本,实现对相似样本差异的有效挖掘。对于给定测试样本,通过集成测试样本多粒度近邻图的分类结果,得到更精准、更鲁棒的智能电表故障预测结果。在20个KEEL(knowledge extraction based on evolutionary learning)和UCI(UC Irvine machine learning repository)不平衡分类公开数据集和智能电表实际故障数据集上的大量实验结果表明,与17种典型方法相比,该文所提算法在处理智能电表故障分类问题上具有显著优势。展开更多
Let AG(n,F q) be the n-dimensional affine space over F q,where F q is a finite field with q elements.Denote by Γ (m) the graph induced by m-flats of AG(n,F q).For any two adjacent vertices E and F of Γ (m)...Let AG(n,F q) be the n-dimensional affine space over F q,where F q is a finite field with q elements.Denote by Γ (m) the graph induced by m-flats of AG(n,F q).For any two adjacent vertices E and F of Γ (m),Γ (m)(E)∩Γ (m)(F) is studied.In particular,sizes of maximal cliques in Γ (m) are determined and it is shown that Γ (m) is not edge-regular when m<n-1.展开更多
密度峰值聚类(clustering by fast search and find of density peaks,DPC)算法在应对大规模聚类时效率不高。k近邻密度支配域小团簇加速技巧可以很好地改善该短板,但存在代表点代表能力不足的问题,从而影响聚类质量。代表团采样策略可...密度峰值聚类(clustering by fast search and find of density peaks,DPC)算法在应对大规模聚类时效率不高。k近邻密度支配域小团簇加速技巧可以很好地改善该短板,但存在代表点代表能力不足的问题,从而影响聚类质量。代表团采样策略可作为上述问题的改进方式。由此形成的新算法不仅继承了原有密度支配域小团簇加速技巧的高效特性,还保证了聚类的质量。算法构建k近邻图。再利用k近邻图进行核密度估计并构建若干个密度支配域。对各密度支配域分别从高低密度区域采样支配域代表团。利用代表团的近邻关系计算域间相似度。将各支配域视为新样本点,执行DPC算法完成聚类。实验证明,引入代表团策略对DPC算法有一定的提升,聚类效果比部分密度聚类算法更好。展开更多
文摘In this paper, we study the flocking behavior of a thermodynamic Cucker–Smale model with local velocity interactions. Using the spectral gap of a connected stochastic matrix, together with an elaborate estimate on perturbations of a linearized system, we provide a sufficient framework in terms of initial data and model parameters to guarantee flocking. Moreover, it is shown that the system achieves a consensus at an exponential rate.
文摘智能电表故障的准确预测对实现计量设备精准主动运维、保障电网稳定运行具有重要意义。电表各故障类型样本的出现频次不同,且不同故障类型样本在高维特征空间中的分布存在重叠,这极大增加了故障预测的难度。现有不平衡分类方法通过构建单一样本信息与其对应类别标签的映射关系来划分样本类型,导致对具有相似表征信息的重叠区样本难以准确判别,降低了整体分类精度。该文提出一种基于多粒度近邻图的智能电表故障分类方法。首先,选择原始数据集中样本作为目标样本,以目标样本及其近邻样本作为节点、目标样本与其近邻样本连线作为边构建近邻图。根据选择的近邻样本数量不同构建多粒度近邻图,实现目标样本的信息扩充和训练样本的数量扩增,更有利于模型稳定训练。构建编码器挖掘近邻图节点特征,利用图注意力机制,根据近邻图节点编码特征和节点邻接关系将近邻样本信息自适应地聚合到目标样本,实现对相似样本差异的有效挖掘。对于给定测试样本,通过集成测试样本多粒度近邻图的分类结果,得到更精准、更鲁棒的智能电表故障预测结果。在20个KEEL(knowledge extraction based on evolutionary learning)和UCI(UC Irvine machine learning repository)不平衡分类公开数据集和智能电表实际故障数据集上的大量实验结果表明,与17种典型方法相比,该文所提算法在处理智能电表故障分类问题上具有显著优势。
基金Supported by the National Natural Science Foundation of China(1 95 71 0 2 4 ) and Hunan Provincial De-partmentof Education(0 2 C5 1 2 )
文摘Let AG(n,F q) be the n-dimensional affine space over F q,where F q is a finite field with q elements.Denote by Γ (m) the graph induced by m-flats of AG(n,F q).For any two adjacent vertices E and F of Γ (m),Γ (m)(E)∩Γ (m)(F) is studied.In particular,sizes of maximal cliques in Γ (m) are determined and it is shown that Γ (m) is not edge-regular when m<n-1.
文摘密度峰值聚类(clustering by fast search and find of density peaks,DPC)算法在应对大规模聚类时效率不高。k近邻密度支配域小团簇加速技巧可以很好地改善该短板,但存在代表点代表能力不足的问题,从而影响聚类质量。代表团采样策略可作为上述问题的改进方式。由此形成的新算法不仅继承了原有密度支配域小团簇加速技巧的高效特性,还保证了聚类的质量。算法构建k近邻图。再利用k近邻图进行核密度估计并构建若干个密度支配域。对各密度支配域分别从高低密度区域采样支配域代表团。利用代表团的近邻关系计算域间相似度。将各支配域视为新样本点,执行DPC算法完成聚类。实验证明,引入代表团策略对DPC算法有一定的提升,聚类效果比部分密度聚类算法更好。