当提供不适当的参数或将其应用于由具有不同形状，大小和密度的聚类组成的数据集时，大多数聚类算法将变得无效。 为了缓解这些不足，我们提出了一种新颖的拆分合并层次聚类方法，其中采用最小生成树（MST）和基于MST的图来指导拆分和合并过程。 在分割过程中，选择基于MST的图中具有高度的顶点作为初始原型，并使用K均值来分割数据集。 在合并过程中，将对子组对进行过滤，并且仅考虑相邻对。 所提出的方法除了簇数以外不需要任何参数。 实验结果证明了其在合成和真实数据集上的有效性。

比较经典的机械振动信号，分为高频、中频和低频，可用在实验中

General perturbations element sets generated by NORAD can be used to predict position and velocity of Earth-orbiting objects. To do this one must be careful to use a prediction method which is compatible with the way in which the elements were generated. Equations for ¯ve compatible models are given here along with corresponding FORTRAN IV computer code. With this information a user will be able to make satellite predictions which are completely compatible with NORAD predictions.

IMM Dataset 是一个人脸数据库，其包括 240 张人脸图片和 240 个 asf 格式文件（可用 UltraEdit 编辑，其中包含 58 个点的地标），每人有 6 张人脸图片，包含 7 名女性和 33 名男性共计 40 人，每张图片均被标记 58 个特征点。 该数据集由丹麦 Tekniske 大学于 2002 年发布。

MUCT Dataset 是一个人脸数据库，主要用于身份鉴定，该数据集由 3755 张人脸图像组成，图片格式为 JPG，其中每张图像包含 76 个点的地标，地标文件分为 CSV、RDA 和 Shape 三种格式，该数据集在种族、年龄等方面表现出极大的多样性。 MUCT Dataset 由开普敦大学于 2010 年发布，主要发布人为 Stephen Milborrow、John Morkel 和 Fred Nicolls，相关论文有《The MUCT Landmarked Face Database》。

详解SQL中Groupings Sets 语句的功能和底层实现逻辑.doc

CS229 吴恩达机器学习 习题大作业答案 problem sets 04 PS04（输出结果语音，欢迎指教）ICA 鸡尾酒晚会语音分解的例子

用rough sets来做属性约简 用matlab编写 多个程序 物有所值

该文档介绍了凸分析中与仿射集相关的基础概念及一些定理，像仿射组合，仿射无关，仿射变换，还有超平面的概念。

A topos is a categorical model of constructive set theory. In particular, the eective topos is the categorical 'universe' of recursive mathematics. Among its objects are the modest sets, which form a set-theoretic model for polymorphism. More precisely, there is a bration of modest sets which satises suitable categorical completeness properties, that make it a model for various polymorphic type theories. These lecture notes provide a reasonably thorough introduction to this body of material, aimed at theoretical computer scientists rather than topos theorists. Chapter 2 is an outline of the theory of brations, and sketches how they can be used to model various typed -calculi. Chapter 3 is an exposition of some basic topos theory, and explains why a topos can be regarded as a model of set theory. Chapter 4 discusses the classical PER model for polymorphism, and shows how it 'lives inside' a particular topos|the eective topos|as the category of modest sets. An appendix contains a full presentation of the internal language of a topos, and a map of the eective topos. Chapters 2 and 3 provide a sampler of categorical type theory and categorical logic, and should be of more general interest than Chapter 4. They can be read more or less independently of each other; a connection is made at the end of Chapter 3. The main prerequisite for reading these notes is some basic category theory: limits and colimits, functors and natural transformations, adjoints, cartesian closed categories. No knowledge of indexed categories or categorical logic is needed. Some familiarity with 'ordinary' logic and typed -calculus is assumed.