Classifcation

Question 1

Kernel Trick

Answer 1

decision boundary can be linear in high-dimensional space & non-linear in the input space

Question 2

K-Fold Cross Validation

Answer 2

1. Create K folds of the data
2. for each fold, use k as the test set and the other folds as the trainings sets.
3. The error is the average of the seperate estimates.
   This is good for small datasets.

Question 3

Leave-one-out cross validation

Answer 3

Like K-Fold but K=dataset size.

Question 4

SVM

Answer 4

• Robust to choice of kernel function

Question 5

Random Forests