I have a simple yes/no problem so I was naturally inclined towards using a binary classifier because I was reading the book, A Course in
Machine Learning by Hal Daumé III and I quote from it:
[ Binary Classification: trying to predict a simple yes/no response.
For instance, predict whether Alice will enjoy a course or not.
Or predict whether a user review of the newest Apple product is
positive or negative about the product.
Multiclass Classification: trying to put an example into one of a number
of classes. For instance, predict whether a news story is about
entertainment, sports, politics, religion, etc. Or predict whether a
CS course is Systems, Theory, AI or Other. ]
But then my professor told me to try using naive Bayes and KNN for my problem which (according to the image I've attached below) are multiclass classifiers.
So can it be beneficial to use multiclass classifiers for a simple binary problem? I'm probably going to try a lot of classifiers to get the best accuracy but before going straight into practical tests I wanted to get some advice because due to the abundance of classifiers I'll have to pick a few(and choose the one providing the best results) and that choice must be based on a certain reasoning.
Best Answer
If you have only two classes it doesn't matter what is the classification rule (one-vs-all or mutually exclusive classes) since for two classes they are the same.