turn on suggestions

Auto-suggest helps you quickly narrow down your search results by suggesting possible matches as you type.

Showing results for

- JMP User Community
- :
- Discussions
- :
- Discussions
- :
- Creating Random distributions

Topic Options

- Subscribe to RSS Feed
- Mark Topic as New
- Mark Topic as Read
- Float this Topic for Current User
- Bookmark
- Subscribe
- Printer Friendly Page

- Mark as New
- Bookmark
- Subscribe
- Subscribe to RSS Feed
- Get Direct Link
- Email to a Friend
- Report Inappropriate Content

Dec 3, 2009 10:35 AM
(1636 views)

Hi,

For teaching purposes, I'm trying to create a dataset with 4 or 5 different distributions that all have a mean of 0 and sigma of 1. The lesson will be that sometimes the variation that is out there is more complicated than can be summarized from just those two numbers.

The first distribution is easy since it will be the "standard" normal distribution where the Random Normal() formula yields mean=0 and sigma=1, but then I get stuck because I haven't been able to figure out to adjust the skewness and kurtosis in the other random number sets. I would also like to add add a few other distribution types like exponential, geometric, and so on.

thanks for the help,

Sean

For teaching purposes, I'm trying to create a dataset with 4 or 5 different distributions that all have a mean of 0 and sigma of 1. The lesson will be that sometimes the variation that is out there is more complicated than can be summarized from just those two numbers.

The first distribution is easy since it will be the "standard" normal distribution where the Random Normal() formula yields mean=0 and sigma=1, but then I get stuck because I haven't been able to figure out to adjust the skewness and kurtosis in the other random number sets. I would also like to add add a few other distribution types like exponential, geometric, and so on.

thanks for the help,

Sean

2 REPLIES

- Mark as New
- Bookmark
- Subscribe
- Subscribe to RSS Feed
- Get Direct Link
- Email to a Friend
- Report Inappropriate Content

I think it will be difficult to achieve that if you restrict the the distribution parameters to mean=0 and sigma=1. Many distributions can't be defined by those parameters. For example, the exponential distribution postulate that mean = sigma (and mean always > 0).

But in principle you may use almost any of jmp's random generators to create random data from different distributions (or just use some real data) and then in other columns use the "col standardize" function to generate new data that by definition have mean =0 and sigma = 1. The standardized data can look skew and kurtotic but still have the properties of mean = 0 and sigma = 1.

But maybe you mean that you want to illustrate that samples taken from populations that follow specific distributions by chance may have a sample average of 0 and a sample standard deviation of one? That might involve a lot of trial and error. Or maybe I am totally missing the point here... (I am no probability expert).

Message was edited by: MS

But in principle you may use almost any of jmp's random generators to create random data from different distributions (or just use some real data) and then in other columns use the "col standardize" function to generate new data that by definition have mean =0 and sigma = 1. The standardized data can look skew and kurtotic but still have the properties of mean = 0 and sigma = 1.

But maybe you mean that you want to illustrate that samples taken from populations that follow specific distributions by chance may have a sample average of 0 and a sample standard deviation of one? That might involve a lot of trial and error. Or maybe I am totally missing the point here... (I am no probability expert).

Message was edited by: MS

- Mark as New
- Bookmark
- Subscribe
- Subscribe to RSS Feed
- Get Direct Link
- Email to a Friend
- Report Inappropriate Content

MS,

I figured it out how to do it using the four-parameter Johnson distribution.

Good anticipation about using a mean of 0. I did change the mean to 10 so I could also use the exponential distribution.

- Sean

I figured it out how to do it using the four-parameter Johnson distribution.

Good anticipation about using a mean of 0. I did change the mean to 10 so I could also use the exponential distribution.

- Sean