If we have two samples (generally their distribution is not known),say $X\sim N(0,1)$, $Y|X\sim N(X,X^2/2)$. Can we recover the conditional CDF of $Y|X$ based on the observed samples in R?


closed as off-topic by Nick Cox, Peter Flom Apr 19 at 11:26

This question appears to be off-topic. The users who voted to close gave this specific reason:

  • "This question appears to be off-topic because EITHER it is not about statistics, machine learning, data analysis, data mining, or data visualization, OR it focuses on programming, debugging, or performing routine operations within a statistical computing platform. If the latter, you could try the support links we maintain." – Nick Cox, Peter Flom
If this question can be reworded to fit the rules in the help center, please edit the question.

  • $\begingroup$ conditional PDF and CDF can be estimated nonparametrically. There is supposedly at least one package available for those purposes. $\endgroup$ – Gary Moore Apr 19 at 5:05
  • $\begingroup$ Thank you for your kind comment! Do you have the reference for the method of nonparametric estimation and the R package? $\endgroup$ – J.Mike Apr 19 at 5:26

You can't determine the CDF from samples, but you can easily get an empirical estimate:

n <- 1000
x <- rnorm(n)
y <- rnorm(n, x, x^2/2)

LM <- lm(y ~ 0 + x) # no intercept because you know this, though usually you won't

Gives me $\hat{\beta} = 1.076$, or $\hat{y} = 1.076 \cdot x + \epsilon$ (about $1\times x$, as you specified).

Similarly, you can get an empirical estimate of the standard deviation you supplied with sd(resid(LM)).

If you don't know anything about their distributions, you could try a non-parametric approach.

  • $\begingroup$ Thank you for your answer! Do you have the reference for the nonparametric approach? $\endgroup$ – J.Mike Apr 19 at 5:28

Finding the conditional distribution of a variable $Y$ conditional on another observed variable $X$ is the entire subject matter of regression analysis (construed in its wide sense to include linear and nonlinear regression models, GLMs, GLMMs, etc.). This is a huge subject and a core part of statistical education. If you would like to learn more about it, I would recommend starting with some material on linear regression analysis, and then building up from there.


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