Introduction Data types and Learning Types

Question 1

Structured data

Answer 1

Numerical data
(age, time, temperature)

Categorical data
(gender, color, country,
class)

Question 2

Unstructured data

Answer 2

• Text
• Audio
• Image
• Signal
• Video

Question 3

Structured data (tree)

Answer 3

Question 4

Tabular data

Answer 4

Columns are features and rows are instances

Question 5

Features

Answer 5

• Features are raw or derived: max, min, average, rank,
bin, etc.

• Time plays a special role: time cannot decrease and
often we want to predict the future based on the
past.

• In case of labeled data, there are descriptive features
and a target feature.

Question 6

Labeled tabular data

Answer 6

Descriptive features

Target feature

Question 7

• Alternative names for descriptive features

Answer 7

• predictor variables
• independent variables

Question 8

• Alternative names for target feature

Answer 8

• response variable
• dependent variable

Question 9

• Alternative names for instances

Answer 9

• individuals, entities,
  cases, objects, or records.

Question 10

Supervised learning using labeled data (goal)

Answer 10

The goal is to find a “rule”

in terms descriptive
features that explains the
target feature as good as
possible.

Question 11

Unsupervised learning (goal)

Answer 11

The goal is to find clusters
or patterns.

• Clusters are
homogeneous sets of
instances.

• Patterns reveal hidden
structures in the data (unknown unknowns)