首页 > > 详细

辅导python Clustering Analysis 程序,辅导R,讲解java AI 编程语言程序辅导R语言程序|解析C/C++编程

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.
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
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.
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.

联系我们 - QQ: 99515681 微信:codinghelp
程序辅导网!