I am from CS background and moving towards Data Sciences, I have came to know ML is highly influenced by Statistical Inference/signal processing. The X we use in data science is called input, feature set or independent variables and Y is called target, class, label. What do you call it? Are they different in market and field among practitioners?

  • $\begingroup$ Thank you for your question. Any chance you can clarify what you mean by the question “what do you call it?”? $\endgroup$ – shepan6 Jul 23 at 18:52
  • $\begingroup$ I mean what are other names for them like I have given a few name there, also found people calling them with other names like in signal processing its different in industrial engineering its different $\endgroup$ – Nauman Akram Jul 23 at 19:04
  • $\begingroup$ Take a look at theory of linear models or simply linear equations: X is the independent variable(s) so most of cases does incorporate a matrix and y( a matrix or a vector so depends on the algorithm) is dependent variable that depends on x. So all this is nothing else than notations used in math and stats is based on Math as we know so no surprise of this notations but is better to keep them this way or you can change the way you want if you want to be different. $\endgroup$ – n1tk Jul 24 at 18:13

A few names I found for X are:

  • feature vector
  • an input variable
  • an independent variable
  • an explanatory variable
  • an exogenous variable
  • a predictor variable
  • a regressor
  • a covariate
| improve this answer | |
  • $\begingroup$ will be glad if someone can share for Y as well $\endgroup$ – Nauman Akram Jul 23 at 19:10
  • $\begingroup$ for Y there is also response variable, dependent variable, sometimes just answer $\endgroup$ – Erwan Jul 23 at 21:57
  • $\begingroup$ It’s worth mentioning that “covariate” can mean other variables besides the one(s) of interest e.g. ANCOVA. In ANCOVA, the group variables are predictor variables, but I’d never call them covariates. Also, “covariate” in this context probably wouldn’t be categorical; such a variable would usually be called a “factor” with “levels”. $\endgroup$ – Dave Jul 24 at 0:11

From personal experience, the majority is inputs or features for X and outputs or labels for Y. Nonetheless, other terms are also common mostly depending on the person's background (some you have already listed).


  • Inputs
  • Features
  • Feature Vector
  • Independent/Explanatory/Exogenous variables
  • Predictor variables


  • Outputs
  • Labels (known outcomes)
  • Dependent/Explained/Predicted variable
  • Outcome
  • Target

Plus, be aware of that target and labels don't always mean the same thing: for some labels represent the known outcomes for the target variable, which means that some inputs may not have a label associated to them (cf. unsupervised/semi-supervised learning).

| improve this answer | |

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