On weighted expectations

The classical 'expected value' frequently fails to represent a 'typical value' of a given multi-cluster data set. In this case the set is clustered and new quantities which replace the expected value and the standard deviation are defined. The 'weighted expected value' (WEV) is computed by solving a nonlinear equation and defines a 'most typical value' of the set. The 'weighted deviation' (WD) is then defined explicitly and provides a measurement to the 'grade of typicality' of the WEV.<
>