Question

Which among the following is an algorithm useful for case based reasoning?

(a) Viterbi algorithm

(b) K- nearest neighbours algorithm

(c) Forward algorithm

(d) Backward algorithm

The question was posed to me during an interview.

This interesting question is from Techniques topic in section Cognitive Techniques of Cognitive Radio

Answer 1

The correct answer is:

(b) K-nearest neighbours algorithm

Explanation:

Case-based reasoning (CBR) is a problem-solving technique that uses previous cases (or experiences) to solve new problems. The K-nearest neighbours (K-NN) algorithm is a well-known method used in CBR because it relies on finding similar cases (or data points) from a stored set of examples, making it a direct fit for this approach.

Here's a breakdown of the options:

• (a) Viterbi algorithm: This is a dynamic programming algorithm primarily used for hidden Markov models (HMMs) to find the most likely sequence of states given a sequence of observed events. It's not typically associated with case-based reasoning.

• (b) K-nearest neighbours algorithm: This algorithm is widely used in machine learning, including case-based reasoning. In K-NN, the algorithm compares a new case with the "K" closest cases in a database to predict an outcome or find a solution. It works by measuring the similarity between cases, making it a good fit for CBR.

• (c) Forward algorithm: The forward algorithm is also used in hidden Markov models to compute the probability of a sequence of observed events. It's not relevant to case-based reasoning.

• (d) Backward algorithm: Similar to the forward algorithm, the backward algorithm is used in hidden Markov models to calculate the likelihood of observing a sequence of events, but it's not related to CBR.

So, K-nearest neighbours is the correct choice because it directly supports the idea of comparing new problems to stored cases to find a solution, which is the core of case-based reasoning.