Sign in

Microsoft.com

Microsoft

Remove From My Forums

Non Gaussian/Emperical Distribution CDF

Question
0

Sign in to vote

If I am working with a data set which is non normal (the tails are way off). Can Infer.NET still be used to handle this? I'd like to be able to obtain/compute the CDF for a bi-variate dataset. How can I do this?

Thursday, August 22, 2013 7:38 PM

All replies

0

Sign in to vote

Please first take a look at http://social.microsoft.com/Forums/en-US/daa215eb-3a08-44c1-b1d7-f7433193af4b/new-distributions-migrated-from-communityresearchmicrosoftcom. If that doesn't help, tell us the model you would like to have, and we could advise on an Infer.NET implementation.

Friday, August 23, 2013 7:44 AM

Owner
0

Sign in to vote
Please first take a look at http://social.microsoft.com/Forums/en-US/daa215eb-3a08-44c1-b1d7-f7433193af4b/new-distributions-migrated-from-communityresearchmicrosoftcom. If that doesn't help, tell us the model you would like to have, and we could advise on an Infer.NET implementation.

Looking more for something like this:

http://www.mathworks.com/help/stats/ksdensity.html

or

http://www.mathworks.com/matlabcentral/fileexchange/19280-bivariant-kernel-density-estimation-v2-1
- Edited by Davewolfs Monday, August 26, 2013 2:29 PM
Monday, August 26, 2013 2:27 PM
0

Sign in to vote
No, Infer.NET does not have kernel density estimation.
- Proposed as answer by John GuiverMicrosoft employee, Owner Thursday, August 29, 2013 8:08 AM
Tuesday, August 27, 2013 1:13 PM

Owner