Precisely, what is a feature? Is it an attribute/property name or its value?

E.g. would features examples be "name", "adress", ...? Or "Dorothy", "123 YellowBrick Road"??

If it is a property name, what do you call its corresponding values? ("Feature-values"?)

If it is a value, how do you call its corresponding property name?

In the texts I read, it seems the meaning is interchangeable and depends on context.


Feature in the data science context is the name of your variable, answering your question it would be things like name, address, price, volume, etc. It is also known as attributes, columns, variables, etc.

Each row in your data set is denominated an instance, in your example again, it would be dorothy, 123 yellowbric road, U$12.3, 1000, etc. They might be called just rows, examples, etc.

I hope it helps.

| improve this answer | |
  • $\begingroup$ Thank you. Now, is there a word for the value taken by a particular feature of an instance (i.e. the "cell" at the intersection of a row (the instance) and a column (the feature)? $\endgroup$ – MonkeyLearning Jul 12 '19 at 18:56
  • $\begingroup$ Not that I am aware of. You can say its a feature value, or an instance of your feature. I think it will be ok. $\endgroup$ – Victor Oliveira Jul 12 '19 at 19:00
  • $\begingroup$ @MonkeyLearning I suppose a more statistical term could be a "realization", which alludes to each feature/column/variable/predictor following an unknown probability distribution and our "observation" being a mere realized value from this distribution. This is kind of pedantic though and I wouldn't worry too much on this. $\endgroup$ – aranglol Jul 12 '19 at 20:53

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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