Community Detection

Question 1

Social Media Data Mining

Answer 1

a method of trying to uncover the hidden patters, trends and other information gathered from a plethora of platforms (Facebook, Twitter and WhatsApp)

Question 2

What are 2 social media data mining techniques?

Answer 2

1) Text Mining
2) Graph Mining

Question 3

Text Mining

Answer 3

the process of extracting information from textual data

Question 4

Graph Mining

Answer 4

extracting information from data represented in graphical form (focuses on community detection)

Question 5

Information

Answer 5

relationships, associations and patterns among the data

Question 6

Data

Answer 6

a collection of facts and statistics to be analyzed on which operations are performed by a computer and can be stored or retrieved

Slide 430

Question 7

Knowledge

Answer 7

information converted to knowledge about historical patterns, which can be used for future predictions

Slide 430

Question 8

What are the data mining use cases?

3

Answer 8

1) Trend Analysis
2) Event Detection
3) Social Spam detection

Question 9

Trend Analysis

Answer 9

Trend analysis is the examination of data over a period to identify patterns, trends, and insights

Slide 431

Question 10

Event Detection (social heat mapping)

Answer 10

Social media heat mapping involves using visual representations (heatmaps) to understand and analyze user interactions on social media platforms.

Slide 433

Question 11

Social Spam Detection

Answer 11

platforms mine social media data to slowly but surely improve at detecting bots

Slide 434

Question 12

What are the characteristics of a social network?

Answer 12

1) They are a collection of entities (like people)
2) relationships are a medium for collecting entities
3) If you have an entity A that is related to entities B and C then you can assume that B and C are related

Question 13

What are the types of social networks?

Answer 13

1) Telephone networks (nodes = phone # & labels = # calls)
2) Email Networks (nodes = email addresses & weak = 1 direction & strong = bidirectional)
3) Collaboration Networks (nodes = indviduals public papers & labels = # papers)

Question 14

Explain the Girvan-Newman Algorithm?

Answer 14

• Edges same level dashed line
• Solid edges = Directed Acrylic Graph (DAG)
• Step 1: Breadth first search
• Step 2: labelling the nodes (root = 1, all other nodes is sum of parents)
• Step 3: Calculate the betweeness (leaf nodes with no child = 1, leaf nodes split between multiple parents split node, other nodes = 1 + sum of credits below)
• # of each edge is the contribution to the betweeness of that edge
• Remove highest betweeness score

Slides 445-447