machine learning week03 exercises

 
题1： 
You are training a classification model with logistic regression. Which of the following statements are true? Check all that apply.

Introducing regularization to the model always results in equal or better performance on examples not in the training set.
Introducing regularization to the model always results in equal or 
better performance on examples not in the training set.
Adding many new features to the modelmakes it likely 同 overfit ont 
the training set. Introducing regularization to the model always 
results in equal or better performance on the training set.
Adding a new feature to the model always results in equal or better 
performance on eamples not in the training set. 
解析： 
答案: 2 
-选项1:more features能够更好的fit 训练集，即新加的feature会提高training set的拟合度,而不是example拟合度。 
-选项2: more features能够更好的fit 训练集,同时也容易导致overfit,正确。 
-选项3: 同1,将正则化方法加入模型并不是每次都能取得好的效果,如果取得太大的化就会导致欠 拟合. 这样不论对traing set 还是 examples都不好. 不正确 。 
-选项4: 新加的feature会提高train set的拟合度,而不是example拟合度. 正确 。 

解析：加入λ会让θ变小，当λ非常非常大时,θ1θ2…θn≈0。所以选择θ值较小的选项。
 
解析： 
答案: 1 
Which of the following statements about regularization are true? Check all that apply.

Using too large a value of λ can cause your hypothesis to dunderfit the 
data.
Using a very large value of λ cannot hurt the performance of your 
hypothesis; the only reason we do not set λ to be too large is to 
avoid numerical problems.
Because regularization causes J(θ) to no longer be convex,gradient descent may not always convege to the globle minimum(when λ>0,and when using an appropriate learing rate a).
Because logistic regression outputs values 0≤hθ(x)≤1, its range of 
output values can only be “shrunk” slightly by regularization anyway, 
so regularization is generally not helpful for it.
选项1: λ太大导致underfit,当λ太大时θ1θ2…θn≈0.只有θ0起作用,拟合出来是一条直线. λ太小才会导致overfit. 正确 *
选项2: 同1. 不正确
选项3: 不正确
选项4: “shrunk” slightly的是θ, regularization是想要解决overfit. 不正确