Copies column values from preceding rows into the current row. The node can be used to
- make a copy of the selected column and shift the cells I steps up ( I = lag interval)
- make L copies of the selected column and shift the cells of each copy 1, 2, 3, ... L -1 steps up ( L = lag = number of copies)
The option "Number of copies" ( L ) in this node is useful for time series prediction. If the rows are sorted in time increasing order, to apply a lag D to the selected column means to place D -1 past values of the column and the current value of the column on one row. The data table can then be used for time series prediction.
The lag interval option I (periodicity or seasonality) in this node is useful to compare values from the past to the current values. Again if the rows are sorted in time increasing order, to apply a lag interval I means to set aside on the same row the current value and the value occurring I steps before.
L and I can be combined to obtain L -1 copies of the selected column, each one shifted I , 2* I , 3* I , ... ( L -1)* I steps backwards.