2
$\begingroup$

For some educational purposes, I need a realistic regression example where the input space is one-dimensional, i.e. $x\in\mathbb{R}$.

I do not want to use any dummy approaches such as simulation from $\sin$ with some noise.

Are you aware of any publicly available data sets?

$\endgroup$
  • $\begingroup$ Fisher's Iris data? $\endgroup$ – DataD'oh Sep 13 '17 at 14:05
  • 1
    $\begingroup$ If you want something more interesting: kaggle.com/datasets simply delete all the features except for one. $\endgroup$ – lnathan Sep 13 '17 at 14:07
  • 1
    $\begingroup$ Fisher's Iris is rather on classification. Isn't it? $\endgroup$ – Karel Macek Sep 13 '17 at 14:10
  • $\begingroup$ Thank you for referring to kaggle, sounds not bad, but I would need an example where the single component is already known. $\endgroup$ – Karel Macek Sep 13 '17 at 14:10
  • 2
    $\begingroup$ Just find any regression dataset and delete all but one of the explanatory variables and the outcome variable... $\endgroup$ – blacksite Sep 13 '17 at 14:43
3
$\begingroup$

For realistic and educational purpose, I may suggest to use Boston data set. It is on UCI repository, and popular in both statistics and machine learning community. In addition, it has "reasonable" amount of rows and columns (~500 observations, and 14 variables).

In R, it is included in MASS package. A staring point can be using housing age and value to do the regression. Intuitively, older house is cheaper.

d=MASS::Boston
fit=lm(medv~age,d)
plot(d$age,d$medv)
abline(fit)

enter image description here

$\endgroup$

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.