Calculating Rolling Kurtosis and Standard Deviation
I have data that is organized by Trial, Timestep and Observation Value. I want to get the rolling kurtosis, standard deviation and skew.
I am currently working with a windows recipe.
I can get Stdev in windows but in the custom aggregations my SKEWNESS and KURTOSIS formulas are failing (I am using the SQL syntax).
Is there a simpler way to achieve what I want? I want the rolling Kurtosis for across all trials at a given time step. For instance at time 25 take the kurtosis of all values observed through time 25 across all scenarios.