Confusion Matrix Flashcards

(27 cards)

2
Q

Confusion Matrix

What types of predictions is a Confusion Matrix used to measure?

aipractitioner

A

Applies to discrete answers, like Classification models (not continuous values)

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

Confusion Matrix

What are the two axes on a confusion matrix?

aipractitioner

A

What the model said vs. what the real world said

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

Confusion Matrix

What is the Accuracy of a model?

aipractitioner

A

The model was right (true positive and true negative)

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

Confusion Matrix

What’s the same about Precision and Recall?

aipractitioner

A

The model said ‘yes’

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
6
Q

Confusion Matrix

What’s happening if the model says ‘yes’ too much?

aipractitioner

A

False positives – bad Precision

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
7
Q

Confusion Matrix

What is the Precision of a model?

aipractitioner

A

How well does a model predict

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
8
Q

Confusion Matrix

What’s the ratio considered for Precision?

aipractitioner

A

true positives vs. false positives

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
9
Q

Confusion Matrix

What is invariant about Precision?

aipractitioner

A

The model said ‘yes’, decide if it was right or not.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
10
Q

Confusion Matrix

What can you minimize by paying attention to Precision?

aipractitioner

A

Minimize the number of false positives

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
11
Q

Confusion Matrix

Example of when high Precision is important?

aipractitioner

A

Saying the defendant is guilty

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
12
Q

Confusion Matrix

What is the Recall of a model?

aipractitioner

A

Does the model say ‘yes’ when it should say ‘yes’?

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
13
Q

Confusion Matrix

What’s the ratio considered for Recall?

aipractitioner

A

True positives vs. false negatives

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
14
Q

Confusion Matrix

What is the invariant about Recall?

aipractitioner

A

The actual value is ‘yes’. What did the model say?

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
15
Q

Confusion Matrix

What’s the hint for remembering “Recall”?

aipractitioner

A

Recall is Real (real answer is ‘yes’, what did the model say?)

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
16
Q

Confusion Matrix

Mental picture for Recall?

aipractitioner

A

“Don’t miss anything” (don’t forget and miss a real true value)

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
17
Q

Confusion Matrix

What can you minimize by paying attention to Recall?

aipractitioner

A

The number of false negatives (times the model should have said ‘yes’, but didn’t)

18
Q

Confusion Matrix

Example of when high Recall is important?

aipractitioner

A

Predicting if you have cancer

19
Q

Confusion Matrix

How can you optimize a model for both Precision and Recall?

aipractitioner

A

You can’t, try to balance the model based on your needs.

20
Q

Confusion Matrix

How do you measure if your model is balanced?

aipractitioner

A

“F1” is a metric that balances Precision and Recall when both are important.

21
Q

Confusion Matrix

What is Model Accuracy?

aipractitioner

A

How well the model guesses correctly

22
Q

Confusion Matrix

Quantitatively, what is Accuracy?

aipractitioner

A

Amount of true positives and true negatives compared to all answers

23
Q

Confusion Matrix

What metric is best when false positives are costly?

aipractitioner

24
Q

Confusion Matrix

What metric is best when false negatives are costly?

aipractitioner

25
Q

Confusion Matrix

Best metric for model that predicts spam email?

aipractitioner

A

Precision (guessed ‘true’ correctly)

26
# *Confusion Matrix* Best metric for model that predicts fraud? | `aipractitioner`
Precision (guessed 'true' correctly)
27
# *Confusion Matrix* Best metric for model that predicts medical tests? | `aipractitioner`
Recall
28
# *Confusion Matrix* What is Specificity? | `aipractitioner`
Ability to correctly identify negatives