Coursera Machine Learning 第八周 quiz Principal Component Analysis

1
point
1. 

Consider the following 2D dataset:

Which of the following figures correspond to possible values that PCA may return for  u(1) (the first eigenvector / first principal component)? Check all that apply (you may have to check more than one figure).

答案AB

1
point
2. 

Which of the following is a reasonable way to select the number of principal components  k ?

(Recall that  n  is the dimensionality of the input data and  m  is the number of input examples.)

答案D

Choose the value of  k  that minimizes the approximation error  1mmi=1||x(i)x(i)approx||2 .

Choose  k  to be 99% of  n  (i.e.,  k=0.99n , rounded to the nearest integer).

Choose  k  to be the smallest value so that at least 1% of the variance is retained.

Choose  k  to be the smallest value so that at least 99% of the variance is retained.

1
point
3. 

Suppose someone tells you that they ran PCA in such a way that "95% of the variance was retained." What is an equivalent statement to this?

答案C

1mmi=1||x(i)x(i)approx||21mmi=1||x(i)||20.05

1mmi=1||x(i)x(i)approx||21mmi=1||x(i)||20.95

1mmi=1||x(i)x(i)approx||21mmi=1||x(i)||20.05

1mmi=1||x(i)x(i)approx||21mmi=1||x(i)||20.95

1
point
4. 

Which of the following statements are true? Check all that apply.

答案BD

Given only  z(i)  and  Ureduce , there is no way to reconstruct any reasonable approximation to  x(i) .

Given input data  xRn , it makes sense to run PCA only with values of  k  that satisfy  kn . (In particular, running it with  k=n  is possible but not helpful, and  k>n  does not make sense.)

PCA is susceptible to local optima; trying multiple random initializations may help.

Even if all the input features are on very similar scales, we should still perform mean normalization (so that each feature has zero mean) before running PCA.

1
point
5. 

Which of the following are recommended applications of PCA? Select all that apply.

答案CD

Preventing overfitting: Reduce the number of features (in a supervised learning problem), so that there are fewer parameters to learn.

To get more features to feed into a learning algorithm.

Data visualization: Reduce data to 2D (or 3D) so that it can be plotted.

Data compression: Reduce the dimension of your data, so that it takes up less memory / disk space.

  • 1
    点赞
  • 1
    收藏
    觉得还不错? 一键收藏
  • 0
    评论

“相关推荐”对你有帮助么?

  • 非常没帮助
  • 没帮助
  • 一般
  • 有帮助
  • 非常有帮助
提交
评论
添加红包

请填写红包祝福语或标题

红包个数最小为10个

红包金额最低5元

当前余额3.43前往充值 >
需支付:10.00
成就一亿技术人!
领取后你会自动成为博主和红包主的粉丝 规则
hope_wisdom
发出的红包
实付
使用余额支付
点击重新获取
扫码支付
钱包余额 0

抵扣说明:

1.余额是钱包充值的虚拟货币,按照1:1的比例进行支付金额的抵扣。
2.余额无法直接购买下载,可以购买VIP、付费专栏及课程。

余额充值