118 views
0 0 votes

Suppose we fit three classifiers and plot the ROC curves for each classifier on the test set. The test set contains 100 points: the first 50 points are labeled 0 and the second 50 points are labeled 1. Determine which models produce each ROC curve.

  1. Predicts $P(Y = 1|X)$ using a random number between 0 and 1
  2. Assigns $P( Y = 1|X) = 0.3$ to the first 50 points and $P( Y = 1|X) = 0.4$ to the second 50 points.
  3. Assigns $P( Y = 1|X) = 0.8$ to the first 50 points and $P(Y = 1|X) = 0.6$ to the second 50 points.
 
  1. ROC A, ROC B, ROC C
  2. ROC A, ROC C, ROC B
  3. ROC B, ROC A, ROC C
  4. ROC A, ROC B, ROC A

Please log in or register to answer this question.

Answer:
Position:
Show:

Related questions

0 0 votes
0 0 answers
97
97 views
GO Classes asked Mar 26, 2025
97 views
A phone manufacturer develops an algorithm to flag each incoming text message as spam or not. A decision of 1 corresponds to spam, and a decision of 0 corresponds to not ...
0 0 votes
1 1 answer
172
172 views
GO Classes asked Mar 26, 2025
172 views
The table below shows a sample of validation data and predictions. Suppose we have 4 classification threshold values: $0, 0.25, 0.5$, and $0.75$. Which ROC $$y_i$$$$1$$$$...
0 0 votes
1 1 answer
126
126 views
GO Classes asked Mar 26, 2025
126 views
Mazi works for a medical device company. He is trying to predict whether manufactured devices will be defective based on photographs of the assembly line. He generates th...
0 0 votes
0 0 answers
100
100 views
GO Classes asked Mar 26, 2025
100 views
We are trying to construct a spam filter for phone calls (an algorithm that classifies calls into spam vs not spam.) The model that we train has the following ROC curve.W...