Ex. 14.1

Ex. 14.1

Weights for clustering. Show that weighted Euclidean distance

\[\begin{equation} d_e^{(w)}(x_i, x_{i'}) = \frac{\sum_{l=1}^pw_l(x_{il}-x_{i'l})^2}{\sum_{l=1}^pw_l}\non \end{equation}\]

satisfies

\[\begin{equation} d_e^{(w)}(x_i, x_{i'}) = d_e(z_i, z_{i'}) = \sum_{l=1}^p(z_{il}-z_{i'l})^2,\non \end{equation}\]

where

\[\begin{equation} z_{il} = x_{il}\cdot \left(\frac{w_l}{\sum_{l=1}^pw_l}\right)^{1/2}.\non \end{equation}\]

Thus weighted Euclidean distance based on \(x\) is equivalent to unweighted Euclidean distance based on \(z\).

Soln. 14.1

By definition of \(z_{il}\) we have

\[\begin{eqnarray} d_e(z_i, z_{i'}) &=& \sum_{l=1}^p \left(\frac{w_l}{\sum_{l=1}^pw_l}\right)(x_{il}-x_{i'l})^2\non\\ &=&\frac{\sum_{l=1}^pw_l(x_{il}-x_{i'l})^2}{\sum_{l=1}^pw_l}\non\\ &=&d_e^{(w)}(x_i, x_{i'}).\non \end{eqnarray}\]