Chapter 10: Machine Learning Flashcards

1
Q

Describe supervised learning

A

Like learning with a teacher; specific data is used to show a desired outcome, such as teaching a formula such as y=mx+b.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
2
Q

Describe unsupervised learning

A

Like pattern recognition without labelled groupings. A real-life example is organizing a set of items - they could be grouped by size, colour, function, or some other commonality.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
3
Q

Why is conditional entropy in the ID3 learning algorithm a good heuristic for selecting an attribute?

A

It determines which attribute has the lowest conditional entropy, and then uses this attribute as the root node of the decision tree. The lowest conditional entropy means less
uncertainty and maximized information gain.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
4
Q

Explain the k-means clustering method

A

Used when we know how many clusters (k) we want
to find. Centroid points are randomly assigned to k points, then each remaining point is clustered to the centroid point closest to it. The centroid points are then recomputed
and the process loops until there is no change in the clusters between consecutive iterations.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
5
Q

Concept learning

A
  • Finding an intention based on an extension (i.e., a set of examples)
  • Express a concept by a set of known concepts
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
6
Q

Relationships between concepts

A
  • Generalization-specialization relation
  • Most general specialization
  • Most specific generalization
How well did you know this?
1
Not at all
2
3
4
5
Perfectly