# pyinterp.fill.loess#

pyinterp.fill.loess(mesh: , nx: int = 3, ny: int = 3, value_type: = None, num_threads: int = 0)[source]#

Filter values using a locally weighted regression function or LOESS. The weight function used for LOESS is the tri-cube weight function, $$w(x)=(1-|d|^3)^3$$.

Parameters:
• mesh – Grid function on a uniform 2-dimensional grid to be filled.

• nx – Number of points of the half-window to be taken into account along the X-axis. Defaults to 3.

• ny – Number of points of the half-window to be taken into account along the Y-axis. Defaults to 3.

• value_type – Type of values processed by the filter. Supported are undefined, defined, all. Default to undefined.

• num_threads – The number of threads to use for the computation. If 0 all CPUs are used. If 1 is given, no parallel computing code is used at all, which is useful for debugging. Defaults to 0.

Returns:

The grid will have NaN filled with extrapolated values.