KURTX.P OpenBI
How heavy the tails of an expression's results over a table's rows are, when they are all there is: the population excess kurtosis.
Excess kurtosis compares the tails with a normal distribution's, which scores 0: positive means more values far out than a bell curve would have, negative fewer (a flat, even spread).
Group: Aggregation. OpenBI’s own: DAX has no function of this name.
Syntax
KURTX.P ( <table>, <expression> )| Parameter | Type | Description |
|---|---|---|
table | Table | The rows to go through. |
expression | Expression evaluated for each row | The number worked out for each row. |
Returns
Number. The population excess kurtosis of the results; BLANK for fewer than two, or when they are all equal.
Remarks
- The population excess kurtosis: n × s4 ÷ s2² − 3, where s2 and s4 add up the squared and fourth powers of the differences from the mean. Excel has no function for it; it is SciPy's
kurtosiswithbias=Trueand type 1 of R's e1071 package. - Rows where the expression is blank are left out; fewer than two results, or results that are all the same, give BLANK.
Compared with DAX
An OpenBI function: DAX has none like it, so a measure or query that uses it works in OpenBI only, not in Power BI or Analysis Services.
Excel has KURT (the sample's, KURTX.S here) but no population kurtosis.
Examples
Eight values
EVALUATE
ROW ( "Excess kurtosis", KURTX.P ( DATATABLE ( "Value", INTEGER, { { 2 }, { 4 }, { 4 }, { 4 }, { 5 }, { 5 }, { 7 }, { 9 } } ), [Value] ) )| Excess kurtosis |
|---|
| -0.21875 |
Two values far apart
The flattest spread there is: half the values at each end.
EVALUATE
ROW ( "Excess kurtosis", KURTX.P ( DATATABLE ( "Value", INTEGER, { { 1 }, { 3 } } ), [Value] ) )| Excess kurtosis |
|---|
| -2 |