10
$\begingroup$

I wonder what the difference between the terms "classifier" and "model" is with relation to classification methodologies for machine learning.

Thanks in advance for your answers!

Improve this question
$\endgroup$
2
  • 1
    $\begingroup$ The term model usually implies some mechanistic insight or logic into how the variables are related (causally or not). The term "classifier" may in some cases refer to a more black-box like system with data as input and labels as output. $\endgroup$
    – drollix
    Commented Aug 18, 2016 at 14:36
  • $\begingroup$ This thread may be of your interest: stats.stackexchange.com/questions/210403/… $\endgroup$
    – Tim
    Commented Aug 18, 2016 at 14:51

3 Answers 3

Reset to default
8
$\begingroup$

I'm definitely no expert in the domain so take my answer with a grain of salt but from what I have understood you have:

  • Classifier : The algorithm, the core of your machine learning process. It can be an SVM, Naive bayes or even a neural network classifier. Basically it's a big "set of rules" on how you want to classify your input.

  • Model : It is what you get once you have finished training your classifier, it's the resulting object of the training phase. You can see it as an "intelligent" black box to whom you feed and input sample and it gives you a label as an ouput.

Hope my answer is clear enough, but yeah the difference is rather subtle between the two terms.

Improve this answer
$\endgroup$
10
$\begingroup$

A classifier is a specific type of model, the output variable of which is discrete, often nominal. As pointed out by others, the terminology is loose.

Improve this answer
$\endgroup$
2
  • 1
    $\begingroup$ This agrees with the introductory material in Pattern Recognition and Machine Learning. $\endgroup$
    – Neil G
    Commented Jul 14, 2019 at 22:25
  • $\begingroup$ I think, this answer is correct and should be accepted. The accepted answer is not quite correct imo. $\endgroup$
    – Andrew Anderson
    Commented Dec 7, 2022 at 15:20
7
$\begingroup$

I don't think there's an unified terminology here, but usually classifier refers to the algorithm to assess classification rules, while the rules themselves is what people often call a model. Otherwise, people call the rules a classifier too, and the algorithms are also referred as models. Also, you can refer to your modelling framework as a model in itself.

Improve this answer
$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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