Parametric approaches:

Here we consider the simple Gaussian model for the two distributions $K_{1}(\mu_{1},\sigma_{1})$ and $K_{2}(\mu_{2},\sigma_{2})$.

• The weighted-Mean-Variance (WMV) is defined as:
  

  $$D(K_{1},K_{2})=\frac{\left\vert\mu_{1}-\mu_{2}\right\vert}{\... ...ac{\left\vert\sigma_{1}-\sigma_{2}\right\vert}{\sigma(\sigma)}$$ (8)

  
  

• Fisher metric is defined as
  
  $$D(K_{1},K_{2})=\frac{\left(\mu_{1}-\mu_{2}\right)^{2}}{\sigma_{1}^{2}+\sigma_{2}^{2}}$$