search

wOOL / InfiGMM

Infinite Gaussian Mixture Model / Variational EM
12 stars 4 forks source link

Clarification about variable names #1

Open ktiwari9 opened 9 years ago

ktiwari9 commented 9 years ago

Hi there, I am interested in Implementing a Dirichlet process mixture of Gaussian processes in matlab. I have my data set which is 57X3 where the format is [Latitude,Longitude,Probability] and I have 57 such readings. I want to use your source code to learn the parameters of DPM-GP by Variational Bayes and cluster the data according to similar probability. How can I do this with your code ?

P.S. In demo file we use parameters N,D,T. I know N is the number of samples we want to draw but what is D and T ?

wOOL commented 9 years ago

Hi,

D is the dimension of feature, in your case, it's 3.

T is the upper bound of number of clusters that the algorithm produces.

Sent from my iPhone

On 22 Oct 2015, at 05:46, Kshitij Tiwari notifications@github.com wrote:

Hi there, I am interested in Implementing a Dirichlet process mixture of Gaussian processes in matlab. I have my data set which is 57X3 where the format is [Latitude,Longitude,Probability] and I have 57 such readings. I want to use your source code to learn the parameters of DPM-GP by Variational Bayes and cluster the data according to similar probability. How can I do this with your code ?

P.S. In demo file we use parameters N,D,T. I know N is the number of samples we want to draw but what is D and T ?

— Reply to this email directly or view it on GitHub.

ktiwari9 commented 9 years ago

I see and how can use my own 57X3 data as an input for this model to perform inference ?