#Sales Offer!| Get upto 25% Off:

# Describe scenarios in which the first method is better than the second and vice versa.

1. Failure of k-fold cross validation Consider a case in that the label is chosen at random according to P[= 1] = P[= 0] = 1/2. Consider a learning algorithm that outputs the constant predictor h(x)= 1 if the parity of the labels on the training set is 1 and otherwise the algorithm outputs the constant predictor h(x)=0. Prove that the difference between the leave-one-out estimate and the true error in such a case is always 1/2.

2. LetH1, . . .,Hk be hypothesis classes. Suppose you are given i.i.d. training examples and you would like to learn the class = ∪ki =1 Hi . Consider two alternative approaches: _ Learn on the examples using the ERM rule _ Divide the examples into a training set of size (1 − α)and a validation set of size αm, for some α ∈ (0,1). Then, apply the approach of model selection using validation. That is, first train each class Hi on the (1−α)training examples using the ERMrule with respect toHi, and let ˆ1, . . .,ˆhk be the resulting\ hypotheses. Second, apply the ERM rule with respect to the finite class {ˆ1, . . .,ˆhk} on the αvalidation examples. Describe scenarios in which the first method is better than the second and vice versa.

#### Found something interesting ?

##### We don't just promise. Here is what we guarantee!

• On-time delivery guarantee
• PhD-level professional writers
• Free Plagiarism Report

• 100% money-back guarantee
• Absolute Privacy & Confidentiality
• High Quality custom-written papers

## Related Model Questions

Feel free to peruse our college and university model questions. If any our our assignment tasks interests you, click to place your order. Every paper is written by our professional essay writers from scratch to avoid plagiarism. We guarantee highest quality of work besides delivering your paper on time.