mkurtosis

Syntax

mkurtosis(X, window, [biased=true], [minPeriods])

Details

Calculate the moving kurtosis of X in a sliding window.

Parameters

biased is a Boolean value indicating whether the result is biased. The default value is true, meaning the bias is not corrected.

Please see mFunctions for the parameters and windowing logic.

Returns

Returns a DOUBLE vector of the same length as the input when the input is a vector.
Returns a matrix with the same dimensions as the input when the input is a matrix.
Returns a table with the same schema as the input when the input is a table.
Returns a tuple with the corresponding structure when the input is a tuple.

Examples

m=matrix(1 9 3 100 3 2 1 -100 9 10000, 1 2 3 4 5 6 7 8 9 100);
m.mkurtosis(8);


#0	#1







3.989653641279048	1.761904761904762
3.989840910744778	1.761904761904762
6.140237905908072	6.101712240467206

m.rename!(date(2020.04.06)+1..10, `col1`col2)
m.setIndexedMatrix!()
mkurtosis(m, 8d)


label	col1	col2
2020.04.07
2020.04.08
2020.04.09	1.5	1.5
2020.04.10	2.3195	1.64
2020.04.11	3.2251	1.7
2020.04.12	4.163	1.7314
2020.04.13	5.1141	1.75
2020.04.14	3.9897	1.7619
2020.04.15	3.9898	1.7619
2020.04.16	6.1402	6.1017

mkurtosis(m, 1w)


label	col1	col2
2020.04.07
2020.04.08
2020.04.09	1.5	1.5
2020.04.10	2.3195	1.64
2020.04.11	3.2251	1.7
2020.04.12	4.163	1.7314
2020.04.13	5.1141	1.75
2020.04.14	3.4937	1.75
2020.04.15	3.4937	1.75
2020.04.16	5.1645	5.145

The default case of kurtosis in DolphinDB is biased (biased = true), while in pandas and Excel it is unbiased estimation, and the kurtosis value 3 of the normal distribution is subtracted.

The following example illustrates the equivalent conversion between the two when using a sliding window:

// python
m = [[1111,2], [323,9], [43,12], [51,32], [6,400]]
df = pandas.DataFrame(m)
y = df.rolling(4).kurt()

// dolphindb
m=matrix(1111 323 43 51 6, 2 9 12 32 400)
m.mkurtosis(4, false)-3


#0	#1



2.504252	2.366838
3.675552	3.941262

Related function: kurtosis