-1

As you see in the figure our textual data need to be classified among some diseases; knowing that a symptom (in the column: label) can be used for multiple diseases.

So, in this case, what type of classification is considered in this problem? Is it a Multi-class classification or a Multi-label classification?

desertnaut
  • 57,590
  • 26
  • 140
  • 166
Siwarr
  • 9
  • 1
  • I’m voting to close this question because it is not about programming as defined in the [help] but about ML theory and/or methodology - please see the intro and NOTE in https://stackoverflow.com/tags/machine-learning/info – desertnaut Jul 28 '22 at 14:51

1 Answers1

0

Multi-Class Classification: There are at least 3 different classes (2 classes = binary classification), e.g. positive, neutral, negative. A sample is assigned with exactly one class:

Sample Class
The weather is nice positive
It is raining heavily, I hate it! negative
I did'nt check the weather neutral

Multi-Label Classification: There are at least 3 different classes (called labels), e.g rain, snow, cold, hot. A sample is assigned with zero, one or multiple labels:

Sample Class
Heavy snowfall was imminent. snow, cold
I did not check the weather
I was a burning summerday hot

In your case, the classes would be the diseases in the column diseases. The column symptoms are used as features for the classification. Each sample (each row) is assigned with exactly one class (one disease). Therefore, it is a multi-class classification.

chefhose
  • 2,399
  • 1
  • 21
  • 32
  • to be more specific the output of the classification in my case would be 1 disease whereas the input data (which are the symptoms) == in the column can be the same for more than one disease but finally one only disease will be the response of the model – Siwarr Jul 29 '22 at 18:09