Could somebody explain this difference just as simple as possible?


Pattern recognition is the “automated discovery of patterns in a training set”, and so it is a general term for machine learning.

Classification is the supervised learning problem whose target value is a finite set of classes (as opposed to regression, wherein the target value is a continuous variable).

Therefore, classification is a kind of pattern recognition problem.

  • $\begingroup$ well stated: I had not considered pattern recognition to be nearly synonymous with ML $\endgroup$ – StephenBoesch Dec 23 '17 at 15:55
  • $\begingroup$ @javadba To be honest, I lifted that from Bishop's introduction in “Pattern recognition and machine learning”. $\endgroup$ – Neil G Dec 23 '17 at 16:35

Classification is simply a more general term than pattern recognition. In both cases, you have a set of classes $K$ and a collection of observations, where each observation is represented by a set of features. Your job is to find a mapping of features to a member of $K$ such that you minimize some measure/estimate of out-of-sample classification error.

In pattern recognition, you are simply using a very complex/large feature space. For example, facial recognition will have an input space equal to the number of pixels in the image. This is no different than classifying a loan application as high or low risk based on several measures of creditworthiness.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.