File Exchange

image thumbnail

Gaussian mixture model parameter estimation with prior hyper parameters

version (3.09 KB) by Rini
Gaussian mixture model using prior hyper parameters based on Expectation Maximization.


Updated 31 Aug 2015

View License

There are quite a few Expectation Maximization based Gaussian mixture models. However, the models do not set any prior for mean and variance. I have implemented a 1D GMM inspired by Chris McCormick. Such a model can be helpful in cases where the data range is small and will prevent kernel overlap by restricting the kernels around the prior values.

Comments and Ratings (1)

najah G

Dear Rini,

can this code been used for n-dimensional where n>1?

MATLAB Release Compatibility
Created with R2014a
Compatible with any release
Platform Compatibility
Windows macOS Linux