Simulating Random Multivariate Correlated Data (Continuous Variables)

Randomly Generated Data Before Cholesky Decomposition

This is a repost of an example that I posted last year but at the time I only had the PDF document (written in \LaTeXe).  I’m reposting it directly into WordPress and I’m including the graphs.

From time-to-time a researcher needs to develop a script or an application to collect and analyze data. They may also need to test their application under a variety of scenarios prior to data collection. However, because the data has not been collected yet it is necessary to create test data. Creating continuous data is relatively simple and is fairly straight forward using the Cholesky (pronounced kol-eh-ski) decomposition. This approach takes an original X variable (or matrix) and uses the Cholesky transformation to create a new, correlated, Y variable. To make things simple and straight forward this example will generate data from the a random normal distribution N(0,1).


The reason this approach is so useful is that that correlation structure can be specifically defined. The scripts can be used to create many different variables with different correlation structures. The method to transform the data into correlated variables is seen below using the correlation matrix R.

R = \left( \begin{smallmatrix} 1&0.8&0.2\\ 0.8&1&0.7\\0.2&0.7&1 \end{smallmatrix} \right)


Once the correlation matrix is set the researcher takes the Cholesky decomposition of the correlation matrix. Multiplying the Cholesky decomposition of the correlation matrix by the data matrix the resulting matrix is a transformed dataset with the specified correlation.

W = \left[Cholesky (R)\right]\left[X\right]
The R code from below will generate a correlation matrix of:

R = \left( \begin{smallmatrix} 1&0.7997999&0.1998661\\ 0.7997999&1&0.7000217\\0.1998661&0.7000217&1 \end{smallmatrix} \right)

Randomly Generated Data After Cholesky Decomposition
R = matrix(cbind(1,.80,.2,  .80,1,.7,  .2,.7,1),nrow=3)
U = t(chol(R))
nvars = dim(U)[1]
numobs = 100000
random.normal = matrix(rnorm(nvars*numobs,0,1), nrow=nvars, ncol=numobs);
X = U %*% random.normal
newX = t(X)
raw =
orig.raw =
names(raw) = c("response","predictor1","predictor2")
plot(head(raw, 100))

Posted in Uncategorized

17 replies on “Simulating Random Multivariate Correlated Data (Continuous Variables)

  1. Thanks, this is very useful. I know I’ve learned this in the past but I’m fuzzy on the details. Do you have a citation or link that provides more details on why this works?

  2. What if I need different distributions on different variables (e.g. first column normal distributed, second one uniform, third one gamma).

    Thanks in advance for your inputs.

    1. When you create the variable ”random.normal <-” you can construct the matrix for each dimension. So rather than having ”nrows=nvars” just set it equal to one (1). Then you can rbind the different distributions together. So that way you can have, for example, a Normal(0,1), a Uniform(1,2), and a Gamma(9,.5). Each will be on a different row of the matrix random.normal.

      1. It works but this is what you get as correlation matrix:
        response predictor1 predictor2
        response 1.00 0.84 0.09
        predictor1 0.84 1.00 0.27
        predictor2 0.09 0.27 1.00

        In fact this was the solution I gave to myself but also the answer comply with my suspects.

      2. could you please post the modified code if i want to generate data from uniform distribution.
        i’m sorry for this request. i’m new in R.
        many thanks

  3. This is great! Do you have any recommendations for what to do when the Cholesky decomposition cannot be computed?

  4. Pls,I want to generate random multivariate data in R for two data sets ie canonical sets using exponential and gamma.Help

  5. hello, sorry for this request.
    i am very new with R. and i want to generate data from uniform distribution using the code.
    could u please post part of the code to modified?
    many thanks

Leave a Reply

Your email address will not be published.