Data Mining (P2)

Question 1

Data mining

Answer 1

The analysis of large datasets to uncover previously unknown patterns, trends or relationships

Question 2

3 pros of data mining

Answer 2

• Helps to understand behaviors and hidden trends/patterns
• Helps in detecting risk and fraud
• Helps analyze large amounts of data very quickly

Question 3

3 cons of data mining

Answer 3

• Can reveal insights that compromise anonymity and privacy
• Expensive
• Requires a large amount of data

Question 4

Cluster analysis

Answer 4

It uses unsupervised learning.

It divides the data into groups (“clusters”) that have similar characteristics without making any assumptions about what the different clusters means.

Question 5

Classification

Answer 5

It uses supervised learning.

A classifier or model is developed and trained with training datasets. When new data is entered the model, a comparison is made against the new data and the known output which is stored in the training datasets. With this comparison, a prediction is made.

Question 6

Association Analysis

Answer 6

It uses unsupervised learning.

Breaks up the data sets into different variables (e.g. gender, age, etc.) and tries to find relationships and dependencies between different variables.

Question 7

Link analysis

Answer 7

It uses unsupervised learning.

Essentially, it establishes relationships/associations between different entities within a data set. It starts off by deciding what is the criteria to create a link between entities, then sees what entities meet the criteria, which do not, and to what extent.

Often used in a criminal situation.

Question 8

Deviation Detection

Answer 8

It uses unsupervised learning.

This data mining technique’s purpose is to detect the anomalies and outliers in a dataset.