skmisc.loess.loess_control¶
- class skmisc.loess.loess_control(surface='interpolate', statistics='approximate', trace_hat='wait.to.decide', iterations=4, cell=0.2)¶
Initialization class for loess control parameters
- Parameters:
- surface
str, optional One of [‘interpolate’, ‘direct’] Determines whether the fitted surface is computed directly at all points (‘direct’) or whether an interpolation method is used (‘interpolate’). The default ‘interpolate’) is what most users should use unless special circumstances warrant.
- statistics
str, optional One of [‘approximate’, ‘exact’] Determines whether the statistical quantities are computed exactly (‘exact’) or approximately (‘approximate’). ‘exact’ should only be used for testing the approximation in statistical development and is not meant for routine usage because computation time can be horrendous.
- trace_hat
str, optional One of [‘wait.to.decide’, ‘exact’, ‘approximate’] Determines how the trace of the hat matrix should be computed. The hat matrix is used in the computation of the statistical quantities. If ‘exact’, an exact computation is done; this could be slow when the number of observations n becomes large. If ‘wait.to.decide’ is selected, then a default is ‘exact’ for n < 500 and ‘approximate’ otherwise. This option is only useful when the fitted surface is interpolated. If surface is ‘exact’, an exact computation is always done for the trace. Setting trace_hat to ‘approximate’ for large dataset will substantially reduce the computation time.
- iterations
int, optional Number of iterations of the robust fitting method. If the family is ‘gaussian’, the number of iterations is set to 0.
- cell
float, optional Maximum cell size of the kd-tree. Suppose k = floor(n*cell*span), where n is the number of observations, and span the smoothing parameter. Then, a cell is further divided if the number of observations within it is greater than or equal to k. This option is only used if the surface is interpolated.
- surface