MATLAB Answers

How to compute AIC using Gaussian Mixture Regression?

8 views (last 30 days)
Joaquim
Joaquim on 25 Apr 2017
Answered: Abhishek Ballaney on 16 Feb 2018
Hello everybody, How can I manually compute the AIC for a Gaussian Mixture Model, with 2 predictor variables and one response variable. The data is approximated by 3 gaussians. I use this ( estimated values already calculated):
n=size(inputVector1,1)+size(inputVector,1);
RSS=sqrt(realValues.^2-estimatedValues.^2);
K=2;
AIC=n.*ln(RSS/n)+2*K;
Can anybody tell me what is wrong? Kind regards Joaquim

  1 Comment

Kiran Javkar
Kiran Javkar on 15 Feb 2018
Your computation of RSS (Residual Sum of Squares) is incorrect. You could use something like:
res = realValues - estimatedValues
RSS = sum(res.^2)
Check the Residuals section under this link: Linear Regression

Sign in to comment.

Answers (1)

Abhishek Ballaney
Abhishek Ballaney on 16 Feb 2018
https://in.mathworks.com/help/stats/tune-gaussian-mixture-models.html
https://in.mathworks.com/help/stats/clustering-using-gaussian-mixture-models.html

  0 Comments

Sign in to comment.

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!