Up to how many different levels can a classifier predict?

Asked

Viewed 30 times

0

I was recently doing a machine learning model and I was wondering if I should use logistic regression. The documentation says that this classifier can predict binary values (0 or 1). In my case, I had 4 possible values and I was in doubt whether I could use it or not. Therefore, I ended up opting for the Forest Random.

However, I was left with a question. If I had used some other classification algorithm, how many values can it predict? In the SVM sample chart, on the scikit-Learn website, it shows up to 3 different possible values.

1 answer

3

Unless it is a binary classifier, there is no limit to the number of classes a classifier can work with. Logistic regression is a binary classifier, so it cannot be used to perform a classification task whose variable response has 3 or more levels.

Like Random Forest, SVM is also multiclass and can therefore be used in the problem in question.

Browser other questions tagged

You are not signed in. Login or sign up in order to post.