Session 6.2

Question 1

Clustering - what is it?

Answer 1

• Finding groups in data.
• Organising data into groups such that there is:
(1) high similarity within each group,
(2) low similarity across the groups.

Question 2

Is clustering the same as classification?

Answer 2

No

• Class labels can be found directly in the data. E.g., blood type.

• Different goals: to “understand” the data better (explore), to organise the
information we have.

Question 3

Distance measures

Answer 3

1. Euclidean distance
2. Manhattan distance
3. Jaccard distance
4. Cosine distance
5. Edit distance (Levenshtein metric)

Question 4

Jaccard distance used when

Answer 4

The possession of a common characteristic between two items is important, but the common absence of a characteristic is not.

• Especially useful when dealing with problems that involve (large) sets of
characteristics that may not be ‘symmetrically’ important.

• Text mining: compare whether two documents contain the same word.

Question 5

Cosine distance often encountered in

Answer 5

text mining or recommendation engines

Question 6

Edit distance (Levenshtein metric)

Answer 6

• Text mining applications.

* Applications: Autocorrect (spelling mistakes).

Question 7

Euclidean distance

Answer 7

• The most common geometric distance measure.
• A numeric dataset with attributes similar in terms of measurement type (similar scale) and units.
• Can be understood as physical distance between two data points.

Question 8

Manhattan distance

Answer 8

• The sum of the absolute differences between pairwise attributes.
• The “taxicab” distance.

Question 9

Cosine distance

Answer 9

The term relates to the method of measurement - the cosine of the angle between two vectors.