Normalization procedure

From CESBA-Wiki
Jump to: navigation, search

Due to the diverse nature of criteria, indicator values are characterized by different unit of measure (i.e. %, Kg, kWh, etc.) and different order of magnitude. For this reason the indicators have to be dimensionalized and rescaled before the aggregation phase.

Normalization consists basically in the assignment of a score to the value or scenario assigned to the indicator. Two main approaches are possible:

  • use of linear performance scales
  • use of performance thresholds

In the case of linear scales the indicator values are normalized in a specific interval. The better the performance, the higher the normalized score. The normalization functions have an upper and lower value. A linear scale awards any improvement of the performance’s level in the specific interval. In the case of thresholds, a specific score is assigned only if the indicators value is above or under a fixed target number. One or more thresholds can be applied to an indicator.

Normalization: linear scale


In the case of thresholds, a specific score is assigned only if the indicators value is above or under a fixed target number. One or more thresholds can be applied to an indicator.


Normalization: single threshold

At the end of the normalization step it is available a set of data composed by the normalized scores associated with each criterion.

The next step is the aggregation procedure.

References

CESBA guide (2014)