【Machine Learning】【Andrew Ng】- Quiz(Week 7)

1、Suppose you have trained an SVM classier with a Gaussian kernel, and it learned the following decision boundary on the training set:

You suspect that the SVM is undertting your dataset. Should you try increasing or decreasing C? Increasing or decreasing sigma?
A. It would be reasonable to try decreasing C. It would also be reasonable to try increasing sigma^2.
B. A. It would be reasonable to try increasing C. It would also be reasonable to try increasing sigma^2.
C. A. It would be reasonable to try decreasing C. It would also be reasonable to try decreasing sigma^2.
D. A. It would be reasonable to try increasing C. It would also be reasonable to try decreasing sigma^2.
答案：D.
underfit，所以lambda过大，需要减小，所以C需要增大，sigma^2需要减小

2、The formula for the Gaussian kernel is given by similarity

The figure below shows a plot of


Which of the following is a plot of

A.

B.

C.

D.

答案：B。
sigma越小，约集中，越瘦。

3、The SVM solves

where the functions cost0(z) and cost1(z) look like this:

The first term in the objective is:

This first term will be zero if two of the following four conditions hold true.
Which are the two conditions that would guarantee that this term equals zero?
A.
B.
C.
D.
答案：BD。
直接代入公式就可得到。

4、Suppose you have a dataset with n = 10 features and m = 5000 examples.
After training your logistic regression classier with gradient descent, you find that it has underfit the training set and does not achieve the desired performance on the training or cross validation sets.
A. Use an SVM with a linear kernel, without introducing new features.
B. Use an SVM with a Gaussian Kernel.
C. Create / add new polynomial features.
D. Increase the regularization parameter λ.
答案：BC.

5、Which of the following statements are true? Check all that apply.
A. Suppose you are using SVMs to do multi-class classification and would like to use the one-vs-all approach. If you have K different classes, you will train K- 1 different SVMs
B. The maximum value of the Gaussian kernel is 1
C. If the data are linearly separable, an SVM using a linear kernel will return the same parameters theta regardless of the chosen value of C (i.e., the resulting value of does not depend on C)
D. It is important to perform feature normalization before using the Gaussian kernel.
答案：BD
A错误，还不知道为什么
B正确，取值范围为0-1.
C.错误
D.正确，一定要记得先标准化