0

A Gaussian mixture model is a weighted sum of Gaussian densities, i.e.,

$L(\theta) = \sum_{i=1}^{m} \pi_{i} f(x_i)$

where $m$ is the number of the mixture component.

Hence, Gaussian mixture models is a sum of a finite mixture Gaussian distribution with unknown parameters. I read that, in Gaussian mixture models, each univariate margins are assumed to be normal. Is that correct and why?

Any help, please?

Maryam
  • 1,012
  • 1
  • 5
  • 20
  • The false premise of this question was comprehensively pointed out the last several times you posted it. Please do not post it again. – Glen_b Aug 28 '18 at 12:00

1 Answers1

2

We do not (need to) assume anything on the margins. Generally, in a mixture model setting we start with a single density $f(\cdot;\theta)$ and then we assume that we have random variables $(X_1, ..., X_N, Z_1, ..., Z_N)$ such that $(X_i, Z_i)$ are independent and for a parameter $\Theta = (\tau_1, ..., \tau_K, \theta_1, ..., \theta_k)$ we have that $X = (X_1, ..., X_N)$ and $Z = (Z_1, ..., Z_N)$ have a common density $p(x,z)$ and that $$p(x|z) = \prod_{i=1}^N p(x_i|z_i) = \prod_{i=1}^N f(x_i;\theta_{z_i})$$ and $$p(z) = \prod_{i=1}^N p(z_i) = \prod_{i=1}^N \tau_{z_i}$$ Let us assume that $N=1$ then we simply compute $$p(x) = \int_{\mathcal{Z}} p(x,z) dz = \int_{\mathcal{Z}} p(x|z)p(z) dz = \sum_{k=1}^K f(x;\theta_{k}) \tau_k$$

i.e. the marginal of each and every data variable $X_i$ is the same and it is this mixture expression that you have written above and not a single Gaussian!

Fabian Werner
  • 3,055
  • 1
  • 9
  • 25
  • thank you so much for your answer. Hence, the margin of each variable is a mixture of Gaussian in the case of the Gaussian mixture model. Is that what you meant by "e. the marginal of each and every data variable Xi X is the same and it is this mixture expression that you have written above and not a single Gaussian!"? – Maryam Aug 28 '18 at 10:00
  • In addition, the margin of each component will be a Gaussian distribution. – Maryam Aug 28 '18 at 10:01