文档下载 文档下载导航
设为首页 | 加入收藏
搜索 请输入内容:  
 导航当前位置: 文档下载 > 所有分类 > 高等教育 > 工学 > Finding a Subset of Representative Points in a Data Set
侵权投诉

Finding a Subset of Representative Points in a Data Set

spectral clustering

IEEE TRANSACTIONS ON SYSTEMS, MAN, AND CYBERNETICS, VOL. 24. NO. 9, SEPTEMBER 1994

1417

. . .

:;.

1 Fig. 1. A density function f on a set[ a, b] .

. . .

. ... . .. .:'. ........... '... .... . .: . . .. .. .. . . . . . . ., . .:, .. ')

. ., . . ..

e .

. . . . . . .... . -' . :E..

,;

I,,

,. .

. .. . .

"

I

. . .: . ..

"

,

.I

.

Count the number of points in S, n (say). Let&=number of representative points

. . . . . . . . .'. . .:,,:,,.. ..: .. . . . .. ... .. . ... ... . . . . . . .. . . . . . . . .. . . . . . .. . . . . . . . '

. ._.

a:.

,

.

Y.,

'

,

"

required

:

.:

. .

. . . . . . . . . 4 ... . . . .: . .

.

1 Find h n

take 8: h, I

-

= Compute ml, m2, and find I$,, in,*,

- - -mn - . . m*,

Fig. 3. A triangular distribution data of size 200. (a) Local best representative point. (b) Cluster center by applying C-means algorithm (C= 1). (c) Almost 5% representative points. (d) Almost 10% representative points. representative points may be considered as the generalization of mode estimation and seed point detection problem. In his classical work[l], Parzen showed how to construct a family of consistent and asymptotically unbiased estimates of probability density function and its mode. The work

was extended by Cacoullos[ 2], among others[9]. In the literature of cluster analysis, there exist several approaches of seed point estimation[8]. Macqueen 131 chooses the first k data units in data set as the initial seed points. Forgy[4] takes any desired partition of the dta units into k mutually exclusive groups and computes the group centroids as seed points. Astrahan[5] computes the density for each data unit as the number of other data units within some specified distance, orders the data units by density, and chooses the one with the highest density as the first seed point. The subsequent seed points are chosen in order of decreasing density, subject to the stipulation that each new seed point is at least a minimum distance away from all other previously chosen seed points. A simpler approach was suggested by Ball and Hall[6]. Here, the overall mean vector of data set is considered as the first seed point. The subsequent seed points are selected by examining the data units in their input sequence and accepting any data unit which is at least some specified distance, say d, from all previously chosen seed points. The process is continued until k seed points are accumulated or the data set is exhausted. Ling[ 111 suggests that each (k,r)-cluster has the property that its elements are within a distance T of at least k other elements of the same cluster, and the entire set can be marked by a chain of links each of length less than or equal to r. But there are no guidelines about k and r. Given n data, the problem is that of finding a small subset of no<< n data that provides good representation of the original data. Let a representative point p represent r data in its neighborhood. Intuitively, p should be such that the sum of its distances from these r data is minimum. But an exact algorithm, which gives the optimal representative point set, is computationally very expensive. See Section VI for details.

find k Find the sets 'SI and find

1=1,2,.-. - - s t

Decrease

I v*

t-l

J

yes

8

Fig. 2. A flowchart of the proposed algorithm AL-2.

Finding a Subset of Representative Points in a Data Set

上一页第2页

你可能喜欢

热门文档

相关文档

  • data mining 9

    each data object is in exactly one subset ? ... – A cluster is a set of points such that ...“representative” point of a cluster 4 center-...

  • NO RESULT

    Analysis is the process of finding a ? subset D’ of D [D: data set to be analyzed...Sampling (determines a representative subset of the...

  • ...FewSparse Modeling for Finding Representative Ob...

    FewSparse Modeling for Finding Representative Objects_文学研究_人文社科_专业资料...few representatives for a dataset, i.e., a subset of data points that ...

  • 第四章 聚类分析

    Finding groups of objects such that the objects ...“representative” point of a cluster 4 center-... which is a subset of the points of the ...

  • chap10_聚类分析

    each data object is in exactly one subset ? ...A cluster is a set of points such that any ...“representative” point of a cluster 4 center-...

  • 数据挖掘第8章

    each data object is in exactly one subset ? ... – A cluster is a set of points such that ...“representative” point of a cluster 4 center-...

  • chap10_聚类分析

    each data object is in exactly one subset ? ...A cluster is a set of points such that any ...“representative” point of a cluster 4 center-...

  • COMP5318 Knowledge Discovery and Data Mining_2011 S...

    ? Finding groups of objects such that the ...“representative” point of a cluster 4 center-...(a) Green points denote the data set in a ...

  • ch10ClusBasic

    ? Finding similarities between data according to ...K-Medoids Clustering: Find representative objects (...points in the test set and the closest centroids...

  • chap3_data_exploration

    ‘gender’ and a representative population of ...data points are summarized in a single figure ?...Selection may involve the chossing a subset of ...

站点地图 | 文档上传 | 侵权投诉 | 手机版
本站所有资源均来自互联网,本站只负责收集和整理,均不承担任何法律责任,如有侵权等其它行为请联系我们.
文档下载 Copyright 2013 doc.guandang.net All Rights Reserved.  -  email