联系方式

  • QQ:99515681
  • 邮箱:99515681@qq.com
  • 工作时间:8:00-23:00
  • 微信:codinghelp

您当前位置:首页 >> CS作业CS作业

日期:2019-10-08 07:23


Visual Clustering Analysis of Social Network (40%)

Social networks are ubiquitous. A fundamental problem related to these networks

is the discovery of clusters or communities. Intuitively, a cluster is a collection of

individuals with dense friendship patterns internally and sparse friendships

externally.

The discovery of close-knit clusters in these networks is of fundamental and

practical interest and the one of major focused problems in AI and Data Mining.

There are many reasons to seek tightly-knit communities in networks, for instance,

target marketing schemes can be designed based on clusters, and it has been

claimed that terrorist cells can be identified.

This assignment gives students two options with different requirements based on

student's disciplinary background, personal preference and existing experience.

This is to satisfy the students who are not major in IT and Mathematics

(Algorithms).

Option One: (group work):

A group of two students are required to work together to analyze an organization's

email network (a type of social networks) through the data clustering and a

clustered graph visualization.

Task 1: through data clustering, we can identify abnormal (implicit) network

patterns that against the hierarchical structure of the organization,

Task 2: through a clustered graph visualization, we can visually read and quickly

understand the data clustering output, including the abnormal network patterns.


Option Two: (individual work):

An individual student is required to visualize an organization's email network (a

type of social networks) through graph visualization and clustered graph

visualization.

Task 1: using graph visualization to visualize attributed email network,

Task 2: using clustered graph visualization to visualize a given clustered email

network that enables readers to quickly understand the data clustering output.

2 | P a g e

The weight of this assignment is 40%.

Specification:

The following Figure shows the organization structure of TACME, sourced from

http://www.tacme.com/corporate_structure.html

A list of staff's ID, Name and Position in TACME

ID Name Position

0 James Director

1 David Director

2 George CEO

3 Ronald Business Development Manager

4 John Business Support Manager

5 Richard Business Control Manager

6 Daniel Sales Department Leader

7 Kenneth Product Department Leader

8 Anthony Marketing Department Leader

9 Robert Project Office Leader

10 Charles Professional Service Leader

11 Paul QA Leader

12 Mark Design Office Leader

13 Kevin Technical Support Office Leader

14 Edward Software Development Leader

15 Joseph Legal Office Leader

16 Michael Finance Office Leader

17 Jason HR Office Leader

3 | P a g e

The Email communication detail in a particular month is shown below:

ID Emails per month Weight ID

0 5 1 1

0 6 1 2

1 5 1 2

2 25 2 3

2 36 2 4

2 53 3 5

3 150 4 6

3 213 5 7

3 298 5 8

4 345 6 9

4 123 4 10

4 212 5 11

4 453 7 12

4 156 4 13

4 278 5 14

5 300 5 15

5 78 3 16

5 256 5 17

6 78 3 7

6 145 4 8

7 139 4 8

9 34 2 10

9 134 4 11

9 546 7 12

9 23 2 13

9 145 4 14

10 256 5 11

10 222 5 12

10 190 4 13

10 56 3 14

11 78 3 12

11 112 4 13

12 98 3 14

15 88 3 16

15 128 4 17

16 238 5 17

17 5 1 7

16 15 2 6

16 23 2 7

4 | P a g e

16 54 3 8

16 18 2 9

16 23 2 11

16 41 2 13

16 13 2 14

16 27 2 10

Weight description:

Quantity Weight

<10 1

11 – 50 2

51 – 100 3

101 – 200 4

201 – 300 5

301 – 400 6

> 401 7

General Requirement:

Option One: (group work)

Students are required:

1) To draw (visualize) the original email network on the paper (or screen) with

the satisfaction of the following Aesthetics Rules: a) Symmetrical Display, b)

Minimization of Edge-Crossings and c) Maximization of Angular Resolution. In

addition, since each edge e in the graph is associated with a weight w(e), you need

to map the w(e) to a graphical attribute, such as color, types of line, size or shapes,

to enhance the readbility of the weight

2) To cluster this email network (or graph) into clustered structures by using

Markov Clustering Algorithm. You need to produce two clustered structures 1)

with the weight w(e), 2) without the weight w(e) .

3) Discuss the findings. If there is one (or more) abnormal network pattern(s)

found, you need to describe them in details.

4) To draw (visualize) these two clustered graphs (one with w(e), another without

w(e) ) on the paper (or screen). Using geometric rectangles (or circles) to bound

clusters in the drawing. Make sure that these regions are not overlapped. In

addition, these drawings shall also satisfy the general graph drawing aesthetics.

5 | P a g e

Option Two: (individual work)

Student is required:

1) To draw (visualize) the original email network on the paper (or screen) with

the satisfaction of the following Aesthetics Rules: a) Symmetrical Display, b)

Minimization of Edge-Crossings and c) Maximization of Angular Resolution. In

addition, since each edge e in the graph is associated with a weight w(e), you need

to map the w(e) to a graphical attribute, such as color, types of line, size or shapes,

to enhance the readability of the weight

2) To draw (visualize) a given clustering of the above email graph, that is: {0, 1, 2},

{3, 4, 5}, {6, 7, 8}, {9, 10, 11}, {12, 13, 14}, {15, 16, 17} on the paper (or screen).

Using geometric rectangles (or circles) to bound clusters in the drawing. Make

sure that these regions are not overlapped. In addition, these drawings shall also

satisfy the general graph drawing aesthetics, including a) Symmetrical Display, b)

Minimization of Edge-Crossings and c) Maximization of Angular Resolution.


版权所有:留学生编程辅导网 2018 All Rights Reserved 联系方式:QQ:99515681 电子信箱:99515681@qq.com
免责声明:本站部分内容从网络整理而来,只供参考!如有版权问题可联系本站删除。