0

Binary relevance is a well known technique to deal with multi-label classification problems, in which we train a binary classifier for each possible value of a feature:

http://link.springer.com/article/10.1007%2Fs10994-011-5256-5

On the other side, one hot encoders (OHE) are commonly used in natural language processing to encode a categorical feature taking multiple values as a binary vector:

http://cs224d.stanford.edu/lecture_notes/LectureNotes1.pdf

Can we consider that these two concepts are the same one? Or are there technical differences?

mountrix
  • 1,126
  • 15
  • 32
  • 1
    They look fairly different to me. Why do you think they are closely related? – Has QUIT--Anony-Mousse Aug 08 '16 at 11:07
  • If you use binary relevance to encode a dataset having a single label per class, it looks like you are applying one-hot encoding on each instance, the vector would be the concatenation of the binary values for all the labels. In multi-target problems, the concepts are different of course. – mountrix Aug 09 '16 at 14:16

1 Answers1

3

Both methods are different.

1. One-Hot encoding

In one-hot encoding, vector is considered. enter image description here

Above diagram represents binary classification problem.

2. Binary Relevance

In binary relevance, we do not consider vector. Following diagram represents class label generation using binary relevance method which is using scalar value.

enter image description here

Sayali Sonawane
  • 12,289
  • 5
  • 46
  • 47