4.第 4 个问题
Which of the following statements are true? Check all that apply.
A 、Given only and Ureduce, there is no way to reconstruct any reasonable approximation to
不正确
B、PCA is susceptible to local optima; trying multiple random initializations may help.
不正确
PCA is a deterministic algorithm: there is no initialization and there are no local optima.
C、input data x∈Rn, it makes sense to run PCA only with values of k that satisfy k<=n(In particular, running it with k=n is possible but not helpful, and k>n does not make sense.)
正确
The reasoning given is correct: with k=n, there is no compression, so PCA has no use.
D、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.
正确
If you do not perform mean normalization, PCA will rotate the data in a possibly undesired way.
???不是很懂B,感觉PCA会有局部最优解,然后记得看的视频里边也说多随机初始化几次会改善这种情况呀