smoothdata

📝Syntax

B = smoothdata(A)
B = smoothdata(A, method)
B = smoothdata(A, method, window)
B = smoothdata(A, dim)
B = smoothdata(A, dim, method)
B = smoothdata(A, dim, method, window)
B = smoothdata(__, nanflag)
B = smoothdata(__, Name, Value)
[B, window] = smoothdata(__)

📥Input Arguments

Parameter	Description
A	input vector or matrix: numeric or logical.
method	a character vector or string: smoothing method. See the description for the list of supported methods.
window	a positive scalar or a two-element vector [back forward]: moving window length.
dim	dimension to operate along: positive integer scalar.
nanflag	a character vector or string: 'omitnan' (default) or 'includenan'.
Name, Value	parameter name/value pairs: 'SamplePoints', 'SmoothingFactor', 'Degree'.

📤Output Arguments

Parameter	Description
B	smoothed data.
window	moving window length used for the smoothing.

📄Description

smoothdata smooths noisy data in a vector or in the columns of a matrix.

By default, smoothdata operates along the first non-singleton dimension using the 'movmean' method and a heuristic window length chosen from the data.

The supported values of method are:

'movmean': moving average over each window (default).

'movmedian': moving median over each window.

'gaussian': moving weighted average with Gaussian weights.

'lowess': local regression using a first-degree polynomial.

'loess': local regression using a second-degree polynomial.

'sgolay': Savitzky-Golay polynomial filter (use 'Degree' to set the polynomial degree, default 2).

The 'omitnan' flag (default) ignores NaN values inside each window, while 'includenan' propagates them.

'SmoothingFactor' is a scalar between 0 and 1 that tunes the automatically chosen window length; larger values give more smoothing.

'SamplePoints' is a vector of uniformly spaced sample coordinates; the window is then expressed in the units of these coordinates.

The robust methods 'rlowess' and 'rloess' are not supported yet.

💡Examples

moving average

A = [1 2 10 4 5];
B = smoothdata(A, 'movmean', 3)

Gaussian smoothing

A = [1 2 10 4 5];
B = smoothdata(A, 'gaussian', 3)

automatically chosen window

A = [1 2 10 4 5];
[B, window] = smoothdata(A)