00:01
Hello students, from the given problem we have to select a gaussian mixture model as a two probability density function for a two -dimensional real -value data synthesis.
00:10
So here we are going to use the gmm with a four component as a two probability density function for two -dimensional real -value data synthesis.
00:19
So here our component are 1, 2, 3, 4 and their corresponding mean vector are 1, 1, 3, 2, 3, 1, 2, 4 and 4, 3 and their covariance metric become 2 .5, 0 .52 and mixing coefficient is 0 .24.
00:40
In this manner we will fill out all the covariance metric for all four factor and it becomes like this with mixing coefficient 0 .25, 2, 1, 1, 2, 0 .25 and the last covariance metric is 1 .5, 0 .5, 0 .51 with mixing coefficient 0 .25.
01:13
Next we have to generate the multiple data set with independently and identically distributed samples using this true -value gmm.
01:22
So here we will generate the four data set with independently identically distributed sample using the true gmm and the data sets will have 10, 100, 1000 and 10 ,000 samples respectively.
01:52
In the next question we have provided that for each data set using the maximum likelihood parameter estimation principle within the framework of k equal to 10 fold cross validation we have to evaluate the gmm with different model order...