Re: How do you find derivatives in noisy data?

"Robert Macy" <macy@xxxxxxxxxxxxxx> writes:

> Is there some simple algorithm that finds the derivative, or second
> derivative, in noisy data?
>
> The data is oversampled which causes the derivative amplitude between
> adjacent data points to be small. This in turn enhances the noise
> differences at each data point to dominate.
>
> Obviously, using two data points further away would help, but then
> throws away all the information between.
>
> Does anybody know how to pull derivatives out of noisy data?

Wiener Filter. An online implementation would be a Kalman filter.

--
David Kastrup, Kriemhildstr. 15, 44793 Bochum
.

Relevant Pages

  • using a cfit result in mathematical operations
    ... I have some noisy data that I could not use for mathematical ... What I did to resolve the problem was to curve fit a smoothing ... derivatives of the curve rather than the "real" data. ... doesn't work for cfit. ...
    (comp.soft-sys.matlab)
  • Re: How do you find derivatives in noisy data?
    ... Fit splines ... > Is there some simple algorithm that finds the derivative, ... > Does anybody know how to pull derivatives out of noisy data? ...
    (sci.math)
  • How do you find derivatives in noisy data?
    ... Is there some simple algorithm that finds the derivative, ... The data is oversampled which causes the derivative amplitude between ... Does anybody know how to pull derivatives out of noisy data? ...
    (sci.math)
  • Re: How do you find derivatives in noisy data?
    ... comparatively small value of n) and doing a linear-regression and using ... the slope of the best-fit line? ... > Does anybody know how to pull derivatives out of noisy data? ...
    (sci.math)
  • Re: How do you find derivatives in noisy data?
    ... > Does anybody know how to pull derivatives out of noisy data? ... Derivative amplifies the noise, so you should filter out the noise as much ... Oversampling should not increase noise, but keep it the same or lower the ...
    (sci.math)