## aggregate.fft()


Aggregate loss `S = X_1 + ... + X_N` by fast Fourier transform.


Usage


``` python
aggregate.fft(
    frequency,
    severity,
    points,
)
```


Works for large claim counts that make [panjer](aggregate.panjer.md#prospicio.aggregate.panjer) underflow. Check `report.aliasing_error` before using the result.


## Parameters


`frequency: (Poisson, NegativeBinomial or Binomial)`  

`severity: Grid`  

`points: int`  


## Returns


`tuple of (Grid, CompoundReport)`  


## Raises


`ValueError`  
If [points](aggregate.CompoundReport.md#prospicio.aggregate.CompoundReport.points) is 0.