密度峰值聚类(clustering by fast search and find of density peaks,简称DPC)是一种基于局部密度和相对距离属性快速寻找聚类中心的有效算法.DPC通过决策图寻找密度峰值作为聚类中心,不需要提前指定类簇数,并可以得到任意形状的簇聚类...密度峰值聚类(clustering by fast search and find of density peaks,简称DPC)是一种基于局部密度和相对距离属性快速寻找聚类中心的有效算法.DPC通过决策图寻找密度峰值作为聚类中心,不需要提前指定类簇数,并可以得到任意形状的簇聚类.但局部密度和相对距离的计算都只是简单依赖基于距离度量的相似度矩阵,所以在复杂数据上DPC聚类结果不尽如人意,特别是当数据分布不均匀、数据维度较高时.另外,DPC算法中局部密度的计算没有统一的度量,根据不同的数据集需要选择不同的度量方式.第三,截断距离dc的度量只考虑数据的全局分布,忽略了数据的局部信息,所以dc的改变会影响聚类的结果,尤其是在小样本数据集上.针对这些弊端,提出一种基于不相似性度量优化的密度峰值聚类算法(optimized density peaks clustering algorithm based on dissimilarity measure,简称DDPC),引入基于块的不相似性度量方法计算相似度矩阵,并基于新的相似度矩阵计算样本的K近邻信息,然后基于样本的K近邻信息重新定义局部密度的度量方法.经典数据集的实验结果表明,基于不相似性度量优化的密度峰值聚类算法优于DPC的优化算法FKNN-DPC和DPC-KNN,可以在密度不均匀以及维度较高的数据集上得到满意的结果;同时统一了局部密度的度量方式,避免了传统DPC算法中截断距离dc对聚类结果的影响.展开更多
为适应数据集分布形状多样性以及克服数据集密度问题,针对已有算法对离群簇检测效果欠佳的现状,提出了一种基于K-近邻树的离群检测算法KNMOD(outlier detection based on K-nearest neighborhood MST)。算法结合密度与方向因素,提出一...为适应数据集分布形状多样性以及克服数据集密度问题,针对已有算法对离群簇检测效果欠佳的现状,提出了一种基于K-近邻树的离群检测算法KNMOD(outlier detection based on K-nearest neighborhood MST)。算法结合密度与方向因素,提出一种基于K-近邻的不相似性度量,然后带约束切割基于此度量构建的最小生成树从而获得离群点。算法可以有效地检测出局部离群点以及局部离群簇,与LOF、COF、KNN及INFLO算法的对比结果也证实了算法的优越性能。展开更多
To resolve the completeness and independence of an invariant set derived by the traditional method, a systematic method for deriving a complete set of pseudo-Zernike moment similarity （translation, scale and rotation...To resolve the completeness and independence of an invariant set derived by the traditional method, a systematic method for deriving a complete set of pseudo-Zernike moment similarity （translation, scale and rotation） invariants is described. First, the relationship between pseudo-Zernike moments of the original image and those of the image having the same shape but distinct orientation and scale is established. Based on this relationship, a complete set of similarity invariants can be expressed as a linear combination of the original pseudo-Zernike moments of the same order and lower order. The problem of image reconstruction from a finite set of the pseudo-Zernike moment invariants （PZMIs） is also investigated. Experimental results show that the proposed PZMIs have better performance than complex moment invariants.展开更多
Study of fuzzy entropy and similarity measure on intuitionistic fuzzy sets (IFSs) was proposed and analyzed. Unlike fuzzy set, IFSs contain uncertainty named hesitance, which is contained in fuzzy membership function ...Study of fuzzy entropy and similarity measure on intuitionistic fuzzy sets (IFSs) was proposed and analyzed. Unlike fuzzy set, IFSs contain uncertainty named hesitance, which is contained in fuzzy membership function itself. Hence, designing fuzzy entropy is not easy because of many entropy definitions. By considering different fuzzy entropy definitions, fuzzy entropy on IFSs is designed and discussed. Similarity measure was also presented and its usefulness was verified to evaluate degree of similarity.展开更多
We investigate the symmetry reduction for the two-dimensional incompressible Navier-Stokes equationin conventional stream function form through Lie symmetry method and construct some similarity reduction solutions.Two...We investigate the symmetry reduction for the two-dimensional incompressible Navier-Stokes equationin conventional stream function form through Lie symmetry method and construct some similarity reduction solutions.Two special cases in [D.K.Ludlow,P.A.Clarkson,and A.P.Bassom,Stud.Appl.Math.103 (1999) 183] and a theoremin [S.Y.Lou,M.Jia,X.Y.Tang,and F.Huang,Phys.Rev.E 75 (2007) 056318] are retrieved.展开更多
Using the machinery of Lie group analysis,the nonlinear system of coupled Burgers-type equations is studied.Using the infinitesimal generators in the optimal system of subalgebra of the said Lie algebras,it leads to t...Using the machinery of Lie group analysis,the nonlinear system of coupled Burgers-type equations is studied.Using the infinitesimal generators in the optimal system of subalgebra of the said Lie algebras,it leads to two nonequivalent similarity transformations by using it we obtain two reductions in the form of system of nonlinear ordinary differential equations.The search for solutions of these systems by using the G'/G-method has yielded certain exact solutions expressed by rational functions,hyperbolic functions,and trigonometric functions.Some figures are given to show the properties of the solutions.展开更多
In this paper, similarity symplectic geometry for curves is proposed and studied. Explicit expressions of the symplectic invariants, Frenet frame and Prenet formulae for curves in similarity symplectic geometry are ob...In this paper, similarity symplectic geometry for curves is proposed and studied. Explicit expressions of the symplectic invariants, Frenet frame and Prenet formulae for curves in similarity symplectic geometry are obtained by using the equivariant moving frame method. The relationships between the Euclidean symplectic invariants, Frenet frame, Frenet formulae and the similarity symplectic invariants, Frenet frame, Frenet formulae for curves are established. Invariant curve flows in four-dimensional similarity symplectic geometry are also studied. It is shown that certain intrinsic invariant curve flows in four-dimensional similarity symplectic geometry are related to the integrable Burgers and matrix Burgers equations.展开更多
文摘密度峰值聚类(clustering by fast search and find of density peaks,简称DPC)是一种基于局部密度和相对距离属性快速寻找聚类中心的有效算法.DPC通过决策图寻找密度峰值作为聚类中心,不需要提前指定类簇数,并可以得到任意形状的簇聚类.但局部密度和相对距离的计算都只是简单依赖基于距离度量的相似度矩阵,所以在复杂数据上DPC聚类结果不尽如人意,特别是当数据分布不均匀、数据维度较高时.另外,DPC算法中局部密度的计算没有统一的度量,根据不同的数据集需要选择不同的度量方式.第三,截断距离dc的度量只考虑数据的全局分布,忽略了数据的局部信息,所以dc的改变会影响聚类的结果,尤其是在小样本数据集上.针对这些弊端,提出一种基于不相似性度量优化的密度峰值聚类算法(optimized density peaks clustering algorithm based on dissimilarity measure,简称DDPC),引入基于块的不相似性度量方法计算相似度矩阵,并基于新的相似度矩阵计算样本的K近邻信息,然后基于样本的K近邻信息重新定义局部密度的度量方法.经典数据集的实验结果表明,基于不相似性度量优化的密度峰值聚类算法优于DPC的优化算法FKNN-DPC和DPC-KNN,可以在密度不均匀以及维度较高的数据集上得到满意的结果;同时统一了局部密度的度量方式,避免了传统DPC算法中截断距离dc对聚类结果的影响.
文摘为适应数据集分布形状多样性以及克服数据集密度问题,针对已有算法对离群簇检测效果欠佳的现状,提出了一种基于K-近邻树的离群检测算法KNMOD(outlier detection based on K-nearest neighborhood MST)。算法结合密度与方向因素,提出一种基于K-近邻的不相似性度量,然后带约束切割基于此度量构建的最小生成树从而获得离群点。算法可以有效地检测出局部离群点以及局部离群簇,与LOF、COF、KNN及INFLO算法的对比结果也证实了算法的优越性能。
基金The National Natural Science Foundation of China(No.61071192,61073138)
文摘To resolve the completeness and independence of an invariant set derived by the traditional method, a systematic method for deriving a complete set of pseudo-Zernike moment similarity （translation, scale and rotation） invariants is described. First, the relationship between pseudo-Zernike moments of the original image and those of the image having the same shape but distinct orientation and scale is established. Based on this relationship, a complete set of similarity invariants can be expressed as a linear combination of the original pseudo-Zernike moments of the same order and lower order. The problem of image reconstruction from a finite set of the pseudo-Zernike moment invariants （PZMIs） is also investigated. Experimental results show that the proposed PZMIs have better performance than complex moment invariants.
基金Project(ER120001) supported by Development of Application Technology BioNano Super Composites, Korea
文摘Study of fuzzy entropy and similarity measure on intuitionistic fuzzy sets (IFSs) was proposed and analyzed. Unlike fuzzy set, IFSs contain uncertainty named hesitance, which is contained in fuzzy membership function itself. Hence, designing fuzzy entropy is not easy because of many entropy definitions. By considering different fuzzy entropy definitions, fuzzy entropy on IFSs is designed and discussed. Similarity measure was also presented and its usefulness was verified to evaluate degree of similarity.
基金Supported by National Natural Science Foundations of China under Grant Nos.10735030,10475055,10675065,and 90503006National Basic Research Program of China (973 Program) under Grant No.2007CB814800+2 种基金PCSIRT (IRT0734)the Research Fund of Postdoctoral of China under Grant No.20070410727Specialized Research Fund for the Doctoral Program of Higher Education under Grant No.20070248120
文摘We investigate the symmetry reduction for the two-dimensional incompressible Navier-Stokes equationin conventional stream function form through Lie symmetry method and construct some similarity reduction solutions.Two special cases in [D.K.Ludlow,P.A.Clarkson,and A.P.Bassom,Stud.Appl.Math.103 (1999) 183] and a theoremin [S.Y.Lou,M.Jia,X.Y.Tang,and F.Huang,Phys.Rev.E 75 (2007) 056318] are retrieved.
文摘Using the machinery of Lie group analysis,the nonlinear system of coupled Burgers-type equations is studied.Using the infinitesimal generators in the optimal system of subalgebra of the said Lie algebras,it leads to two nonequivalent similarity transformations by using it we obtain two reductions in the form of system of nonlinear ordinary differential equations.The search for solutions of these systems by using the G'/G-method has yielded certain exact solutions expressed by rational functions,hyperbolic functions,and trigonometric functions.Some figures are given to show the properties of the solutions.
基金supported by National Natural Science Foundation of China(Grant Nos.11471174 and 11101332)Natural Science Foundation of Shaanxi Province(Grant No.2014JM-1002)the Natural Science Foundation of Xianyang Normal University of Shaanxi Province(Grant No.14XSYK004)
文摘In this paper, similarity symplectic geometry for curves is proposed and studied. Explicit expressions of the symplectic invariants, Frenet frame and Prenet formulae for curves in similarity symplectic geometry are obtained by using the equivariant moving frame method. The relationships between the Euclidean symplectic invariants, Frenet frame, Frenet formulae and the similarity symplectic invariants, Frenet frame, Frenet formulae for curves are established. Invariant curve flows in four-dimensional similarity symplectic geometry are also studied. It is shown that certain intrinsic invariant curve flows in four-dimensional similarity symplectic geometry are related to the integrable Burgers and matrix Burgers equations.
