孤立点

In order to overcome this difficulty,a clustering algorithm based on dynamic division of connected graph is proposed.

当前大部分的聚类算法都难以处理任意形状和大小、存在孤立点和噪音以及密度多变的簇，为此，文中提出了一种基于连通图动态分裂的聚类算法。

Some valuable works in the paper include: First, based on the algorithm of based-cell detection, the paper presented a dynamic adjustment function on "M" which can work out a problem of outliers fault-making on boundary cell.

总结本文的研究内容，主要有如下创新： 1、在基于单元的孤立点数据挖掘算法基础上，针对边界单元格出现的孤立点误判问题，给出了对M值进行动态调整函数。

Theorem4.5 Let A be a compact normal operator, is an isolated point in Prop4.13 Let A be a compact normal operator, AO is an isolated point in a.

i20 其中（入，…＼＊为一川的所有的互不相同的孤立点。命题4.13设A是紧正常算子，人。一0是a…）的＄良立点。

Let Kv be a complete graph with v vertices, G be a simple subgraph without isolate vertices of Kv.

设Kv是一个v个点的完全图， G为一个不含孤立点的简单图。

Using traditional supervised and unsupervised classification approaches, combining wiping off isolated point and isolated area, I want to check out the changed areas.

吸收传统变化检测方法中的监督、非监督分类方法，结合分类后孤立点和孤立区域的去除，检测出有变化的区域。

What's more, we also study some key techniques and realization difficulties of image authentication in-depth, including referent block to help tamper localizing, sign bit adjustment for binary encoding, mean quantization improvement, diagonally extension to help tamper localizing, and discard isolated error point to tolerate incidental distortion. These techniques can be used for good reference to the development of image authentication technique.

此外，我们还对数字图像鉴别的一些关键技术和实现难点进行了深入研究，主要包括有：参考点辅助定位技术、二值编码中的符号处理技术、均值量化方法的改进技术、对角扩展和去除孤立点技术等，这些技术的研究都对数字图像鉴别技术的发展具有很好的借鉴意义。

This algorithm uses the gird and density threshold method to remove isolated points of data set,selects the initial seed points in propagating tree randomly,judges vegetal direction and executes sort according to the density-based distance,calculates cluster aim function by the propagating tree value.

算法利用网格和密度阈值去除数据集中的孤立点，从网格集中随机地选取种子点，以基于密度距离作为判断生长方向及分类的依据，以网格生长树的大小作为聚类目标函数。

In this paper, firstly, not only the incidence matrix ,adjacent matrix, cycle matrix, cut-set matrix of an undirected graph are summarized, but also the close contact between a graph and its corresponding matrix are discussed ; secondly, many problems of a graph which are solved by analysing its matrix are listed as follows:1、The co-tree set of a graph is obtained by using its cycle-matrix ; 2、The branches of its spanning tree are given by using its cut-set matrix ; 3、By making use of the incidence matrix of a graph ,not only its vertex cut 、cut vertex 、isolated point and spanning tree can be obtained ,but also the two sides which are whether parallel or not can be judged ;4、By using their adjacent matrix ,the two graphes which are whether isomorphous or not can be judged; once more, there is a detailed introduction in view of special graph (for example: bigaritite graph ,regular graph and so on);last but not least, a graph method of calculating the N power of a matrix is given and the practical applications of the theorem for degree is indicated.

本文首先综述了无向图的关联矩阵，邻接矩阵，圈矩阵，割集矩阵以及图和它对应矩阵之间的关系；其次总结出了利用上述各类矩阵可以解决的图的若干问题：1、利用图的圈矩阵可以求其连枝集；2、利用图的割集矩阵可以求其生成树的树枝；3、利用图的关联矩阵不仅可以求其割点、点割集、连通度、孤立点和生成树，而且可以判断两条边是否平行；4、利用图的邻接矩阵可以判断两个图是否同构；再次，针对特殊图（例如：二分图、正则图等等）的邻接矩阵作了详细介绍；最后，得到了利用图计算矩阵的N次幂的方法，指出度数定理的实际应用。

We led the density function into the data structure as the weight, and use dense micro-cluster to describe the arbitrary-shaped clusters in data streams. In addition, we propose candidate core-micro-cluster and isolated micro-cluster synopses to maintain and distinguish the potential micro clusters and isolation point in data streams.

我们把密度函数以权值的形式引入数据结构中，并利用核心微簇描述数据流中任意形状的簇，并提出候选核心微簇和孤立微簇结构分别用于维护并区分数据流中潜在的核心簇和孤立点。

This new strategy gets the potential support vectors and removes the non-boundary outlier data immixed in the other class by using fuzzy clustering. That can greatly reduce the scale of the training sample set and improve the generalization performance by effectively avoiding over-learning caused by outlier data, and finally speed up learning rate without reducing the classification accuracy.

该策略运用模糊聚类方法快速的提取出潜在支持向量并去除类内非边界孤立点，在减小训练样本集规模的同时，能够有效地避免孤立点数据所造成的过学习现象，提高了SVM的泛化性能，在保证不降低分类精度的前提下提高训练速度。

This one mode pays close attention to network credence foundation of the businessman very much.

这一模式非常关注商人的网络信用基础。

Cell morphology of bacterial ghost of Pasteurella multocida was observed by scanning electron microscopy and inactivation ratio was estimated by CFU analysi.

扫描电镜观察多杀性巴氏杆菌细菌幽灵和菌落形成单位评价遗传灭活率。