Question

What is instance based learning?

(a) It is a family of algorithms that generates new problem instances

(b) It is a family of algorithms that compares new problem instances with instances available in training

(c) It is a family of algorithms that compares new problem instances with each other to determine the degree of similarity

(d) It is a family of algorithms that compares new problem instances with each other to determine the degree of dissimilarity

I got this question in class test.

I want to ask this question from Architecture topic in section Cognitive Radio Architecture of Cognitive Radio

Answer 1

Right option is (b) It is a family of algorithms that compares new problem instances with instances available in training

To elaborate: Instance based learning is a family of algorithms that compares new problem instances with instances available in training. Instance based learning builds hypotheses directly from training instances. The significance of instance based learning is that hypotheses complexity increases with increase in data.