What should I do with NaN

First time being in the forum...

I have a set of EMG data (with a column of time and a column of corresponding EMG data). Unfortunately, there were missing data points during the data collection(so those data do not exist in my data set) and so I substituted those point with NaN.

Here's the problem: if I want to filter the EMG data, it will give out a column of NaN because my original vector contains NaN. So what should I substitute with the missing EMG data point instead?

Thank you

Answer 1

[Walter Roberson]

You could interpolate the signal at the points that are NaN.

One convenient routine for that is John D'Errico's File Exchange contribution inpaint_nan