Standard deviance(sd)

aggregation-functions standard deviance

syntax

  • sd(a)
  • sd(a, relation)

definition

  • sd(a) results in a parameter with the standard deviance of the non null values of attribute a.
  • sd(a, relation) results in an attribute with the standard deviance of the non null values of attribute a, grouped by relation. The domain-unit of the resulting attribute is the values-unit of the relation.

description

The standard deviation is calculated as the square of the sum of the differences of each value related to the mean value. This square is divided by the number of values (n) and from this division the square root is taken.

This way of calculating the standard deviation is different from e.g. the ms Excel STDEV function, which divides the sum of the differences by n -1.

applies to

  • attribute a with Numeric value-type
  • relation with value type of the group CanBeDomainUnit

conditions

  1. The value type of argument a and the resulting data-item must match.
  2. The domain of arguments a and relation must match.

example 

1. parameter<float32> sdNrInh                := sd(float32(City/NrInhabitants)); result = 155.46
2. attribute<float32> sdNrInhRegion (Region) := sd(float32(City/NrInhabitants), City/Region_rel);
City/NrInhabitants City/Region_rel
550 0
525 1
300 2
500 1
200 3
175 null
null 3

domain City, nr of rows = 7

sdNrInhRegion
0
12.5
0
0
null

domain Region, nr of rows = 5

see also