This question is only about the vocabulary.

Do / can you say

  1. data item
  2. data sample
  3. recording
  4. sample
  5. data point
  6. something else

when you talk about elements of the training / test set? For example:

The figure shows 100 data items of the training set.

Database A contains the same data items as database B, but in another format.

The remaining data items were removed from the dataset.

Those 10 classes have 123456 data items.

Please provide papers with examples.

According to Google n-grams:

enter image description here

  • 1
    $\begingroup$ "datum" is what I would use $\endgroup$
    – user4710
    Commented Jan 19, 2017 at 17:02
  • $\begingroup$ If you're going to talk about training/set sets you can simply call them elements, following set theory terminology. $\endgroup$
    – Emre
    Commented Jan 19, 2017 at 17:23
  • 1
    $\begingroup$ Then you could refer to bijections, injections, or surjections, since they're not really the same (if one is a crop of another), but related. $\endgroup$
    – Emre
    Commented Jan 19, 2017 at 18:46
  • 1
    $\begingroup$ In my opinion, it depends on the field you are working in. In my experience, after reading many papers in some specific machine learning and time-series topics (mostly hidden Markov models), I've almost never seen "data item" but most of the time "data samples" or "data points". This led me to use the same vocabulary in my publications. I guess that from the results you provide, other specific fields of study would rather use "data item". Let's remember that the essential is to be understood so, sticking to the vocabulary of the field is for me the rule. :) $\endgroup$
    – Eskapp
    Commented Jan 19, 2017 at 21:28
  • 1
    $\begingroup$ I would call them "data points" or "observations." A "sample" sounds like a lot of data points. $\endgroup$ Commented Jan 19, 2017 at 21:58

1 Answer 1


The term you are looking for is "Example". Source: Martin Zinkevich, Research Scientist at Google (http://martin.zinkevich.org/rules_of_ml/rules_of_ml.pdf)

Instance: The thing about which you want to make a prediction. For example, the instance might be a web page that you want to classify as either "about cats" or "not about cats".

Label: An answer for a prediction task ­­ either the answer produced by a machine learning system, or the right answer supplied in training data. For example, the label for a web page might be "about cats".

Feature: A property of an instance used in a prediction task. For example, a web page might have a feature "contains the word 'cat'".

Example: An instance (with its features) and a label.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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