I need to discuss the systematic component of the model in bivariate linear regression, but what is it in the first place? I have never come across this terminology in our class textbooks.

  • 3
    $\begingroup$ See stats.oecd.org/glossary/detail.asp?ID=3861. $\endgroup$ – Scortchi - Reinstate Monica Jan 17 '16 at 15:19
  • $\begingroup$ @Scortchi - and this is why I give some hints etc in the comments. I see the "graybeards" of the subject leading. $\endgroup$ – EngrStudent Feb 6 '16 at 1:47
  • $\begingroup$ "Bivariate linear regression" is potentially ambiguous. Do you mean (i) a single-x-single-y situation, (ii) a situation where $Y$ is a vector of length two, or (iii) a situation where there are two predictors (Y~X1+X2)? $\endgroup$ – Glen_b -Reinstate Monica Feb 6 '16 at 5:38

Many models for data can be conceived as having the form

$$E(Y|\,\underline{X}=\underline{x}\,\!) = \mu(\,\!\underline{x}\,\!)+\epsilon$$

where given $x$, $\mu(x)$ is fixed; this is called the systematic component of the data-generating process, while $\epsilon$ is the random component.

We're often interested in estimating $\mu$ (as a function of $x$) and some characteristics of $\epsilon$ (such as its variance).


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.