1R

Question 1

Difference between classification and association rules?

Answer 1

Classification rules can only be used to predict the class, no other attribute whereas association rules can predict any attributes value not just the class and multiple attributes value at a time

Question 2

How does 1R deal with numeric attrbutes?

Answer 2

Discretize (divide each attributes range into intervals)

Question 3

How does 1R deal with overfitting?

Answer 3

Overfitting: noise will probably produce a seperate interval.

Set a minimum number of instances in the majority class per interval

Question 4

What does 1R do?

Answer 4

for each attribute, count the different values for the class and you count the errors and choose the ones with minimal errors.

uses one attribute