Prev: Replacing NaN Values...?
Next: difficulties in generating variable PWM
From: Neha on 18 Jun 2010 05:16
Hi,

I was doing ID Interpolation using Spline in order to fill my NaN values. But the result is showing some negative values which is not correct. I thought of converting them to zero but then that might be wrong to do since there are many such cases.

Can anyone suggest, how this can be removed or analysed further. As per my prliminary analysis the data I have is right skewed, so can I do something for that intially and then compute the Nan values.

Any help will be great.

Thanks
Neha
From: John D'Errico on 18 Jun 2010 06:02
"Neha " <smile_neha16(a)yahoo.co.in> wrote in message <hvfdgl$kuq$1(a)fred.mathworks.com>...
> Hi,
>
> I was doing ID Interpolation using Spline in order to fill my NaN values. But the result is showing some negative values which is not correct. I thought of converting them to zero but then that might be wrong to do since there are many such cases.
>
> Can anyone suggest, how this can be removed or analysed further. As per my prliminary analysis the data I have is right skewed, so can I do something for that intially and then compute the Nan values.
>

The simplest answer is to use pchip for your interpolant.

John
From: Neha on 18 Jun 2010 06:21
Okays. Thanks John, I will try with pchip.

Neha

"John D'Errico" <woodchips(a)rochester.rr.com> wrote in message <hvfg6t$9a2$1(a)fred.mathworks.com>...
> "Neha " <smile_neha16(a)yahoo.co.in> wrote in message <hvfdgl$kuq$1(a)fred.mathworks.com>...
> > Hi,
> >
> > I was doing ID Interpolation using Spline in order to fill my NaN values. But the result is showing some negative values which is not correct. I thought of converting them to zero but then that might be wrong to do since there are many such cases.
> >
> > Can anyone suggest, how this can be removed or analysed further. As per my prliminary analysis the data I have is right skewed, so can I do something for that intially and then compute the Nan values.
> >
>
> The simplest answer is to use pchip for your interpolant.
>
> John
 | 
Pages: 1
Prev: Replacing NaN Values...?
Next: difficulties in generating variable PWM