MATLAB: Which interpolation method is best

timeseries

Hi,

I have a colomn of temperature data in a 4 column width table.

Some of the temp data has NANs – but it is regulary spaced over time.

I want to interpolate the temp data (to fill in the NANs) 4 times using different methods each time

linear
spline
pchip
makima

I can calculate the variance of each interpolation and compare it to the variance of the original data

The interpolation with the most similar variance will be selected.

Can you suggest code to do this?

Thank you

Best Answer

Use https://www.mathworks.com/help/matlab/ref/fillmissing.html#bva1z1c-method fillmissing() -- it supports all of the methods you are interested in.

Related Solutions

MATLAB: Transfer function which is the best

Always use the default. There is no need to specify. See the documentation for an alternative only if you are having a problem caused by this choice.

For regression and curvefitting,

 help fitnet
 doc fitnet
 help trainlm
 doc trainlm

For tiny data sets trainbr works better

For huge data sets trainrp and trainscg work better

For classification and pattern recognition

 help patternnet
 doc patternnet
 help trainscg
 doc trainscg

Hope this helps.

Thank you for formally accepting my answer

Greg

MATLAB: Interpolation method (spline)

From the documentation:

'spline':

‘Spline interpolation using not-a-knot end conditions. The interpolated value at a query point is based on a cubic interpolation of the values at neighboring grid points in each respective dimension.’

It does not appear to be the same as 'cubic'. According to the documentation, 'cubic' is currently the same as 'pchip'.

Best Answer

Related Solutions

MATLAB: Transfer function which is the best

MATLAB: Interpolation method (spline)

Related Question