Thread Subject: svd and nan

Hello,

I need to use the bart test (statistical tool box). For
that I need to create a matrix, each column represent a
group of measurement, but i don't have the same
measurements number in each group, so I've created a nan
matrix. The only problem is that the svd m file doesn’t
know what to do with nan. Anyone has an idea for me?

thanks,
Buffy

On 26 Nov, 12:50, "Buffy Summers" <off...@gmail.com> wrote:
> Hello,
>
> I need to use the bart test (statistical tool box). For
> that I need to create a matrix, each column represent a
> group of measurement, but i don't have the same
> measurements number in each group, so I've created a nan
> matrix. The only problem is that the svd m file doesn’t
> know what to do with nan. Anyone has an idea for me?

Use something else than nan for your invalid data point.
Or truncate all the data sets to the same length as the
shortest data set.

Rune

Hello

I can't use a number because it will affect my results. I
can't truncate it as well.
Any other idea?

Thanks,
Rune Allnor <allnor@tele.ntnu.no> wrote in message
<18a18001-5547-417d-9172-
4740bee871f8@d4g2000prg.googlegroups.com>...
> On 26 Nov, 12:50, "Buffy Summers" <off...@gmail.com>
wrote:
> > Hello,
> >
> > I need to use the bart test (statistical tool box). For
> > that I need to create a matrix, each column represent a
> > group of measurement, but i don't have the same
> > measurements number in each group, so I've created a nan
> > matrix. The only problem is that the svd m file doesn’t
> > know what to do with nan. Anyone has an idea for me?
>
> Use something else than nan for your invalid data point.
> Or truncate all the data sets to the same length as the
> shortest data set.
>
> Rune

"Buffy Summers" <offdar@gmail.com> schrieb im Newsbeitrag
news:fiedv7$292$1@fred.mathworks.com...
> Hello
>
> I can't use a number because it will affect my results. I
> can't truncate it as well.
> Any other idea?
>
> Thanks,
> Rune Allnor <allnor@tele.ntnu.no> wrote in message
> <18a18001-5547-417d-9172-
> 4740bee871f8@d4g2000prg.googlegroups.com>...
>> On 26 Nov, 12:50, "Buffy Summers" <off...@gmail.com>
> wrote:
>> > Hello,
>> >
>> > I need to use the bart test (statistical tool box). For
>> > that I need to create a matrix, each column represent a
>> > group of measurement, but i don't have the same
>> > measurements number in each group, so I've created a nan
>> > matrix. The only problem is that the svd m file doesn’t
>> > know what to do with nan. Anyone has an idea for me?
>>
>> Use something else than nan for your invalid data point.
>> Or truncate all the data sets to the same length as the
>> shortest data set.
>>
>> Rune
>

Hi,

what about interpolation on a common time stamp?

Titus

Hi,

Interpolation is not possible.

I don't know what this svd does. Is it possible to change
the array of the matrix to the way it's done in the ANOVA?
i.e. one column is the data and the other column is the
affiliation (group number).

Thanks,
Buffy

"Titus" <titus.edelhofer@mathworks.de> wrote in message
<fiehkc$oro$1@fred.mathworks.com>...
>
> "Buffy Summers" <offdar@gmail.com> schrieb im Newsbeitrag
> news:fiedv7$292$1@fred.mathworks.com...
> > Hello
> >
> > I can't use a number because it will affect my results.
I
> > can't truncate it as well.
> > Any other idea?
> >
> > Thanks,
> > Rune Allnor <allnor@tele.ntnu.no> wrote in message
> > <18a18001-5547-417d-9172-
> > 4740bee871f8@d4g2000prg.googlegroups.com>...
> >> On 26 Nov, 12:50, "Buffy Summers" <off...@gmail.com>
> > wrote:
> >> > Hello,
> >> >
> >> > I need to use the bart test (statistical tool box).
For
> >> > that I need to create a matrix, each column
represent a
> >> > group of measurement, but i don't have the same
> >> > measurements number in each group, so I've created a
nan
> >> > matrix. The only problem is that the svd m file
doesn’t
> >> > know what to do with nan. Anyone has an idea for me?
> >>
> >> Use something else than nan for your invalid data
point.
> >> Or truncate all the data sets to the same length as the
> >> shortest data set.
> >>
> >> Rune
> >
>
> Hi,
>
> what about interpolation on a common time stamp?
>
> Titus
>
>

"Buffy Summers" <offdar@gmail.com> wrote in message
<fiekfg$3sc$1@fred.mathworks.com>...
> Hi,
>
> Interpolation is not possible.
>
> I don't know what this svd does. Is it possible to change
> the array of the matrix to the way it's done in the ANOVA?
> i.e. one column is the data and the other column is the
> affiliation (group number).
>
> Thanks,
> Buffy
>
> "Titus" <titus.edelhofer@mathworks.de> wrote in message
> <fiehkc$oro$1@fred.mathworks.com>...
> >
> > "Buffy Summers" <offdar@gmail.com> schrieb im Newsbeitrag
> > news:fiedv7$292$1@fred.mathworks.com...
> > > Hello
> > >
> > > I can't use a number because it will affect my results.
> I
> > > can't truncate it as well.
> > > Any other idea?
> > >
> > > Thanks,
> > > Rune Allnor <allnor@tele.ntnu.no> wrote in message
> > > <18a18001-5547-417d-9172-
> > > 4740bee871f8@d4g2000prg.googlegroups.com>...
> > >> On 26 Nov, 12:50, "Buffy Summers" <off...@gmail.com>
> > > wrote:
> > >> > Hello,
> > >> >
> > >> > I need to use the bart test (statistical tool box).
> For
> > >> > that I need to create a matrix, each column
> represent a
> > >> > group of measurement, but i don't have the same
> > >> > measurements number in each group, so I've created a
> nan
> > >> > matrix. The only problem is that the svd m file
> doesn’t
> > >> > know what to do with nan. Anyone has an idea for me?
> > >>
> > >> Use something else than nan for your invalid data
> point.
> > >> Or truncate all the data sets to the same length as the
> > >> shortest data set.
> > >>
> > >> Rune
> > >
> >
> > Hi,
> >
> > what about interpolation on a common time stamp?
> >
> > Titus
> >
> >
>

Hi there,
did you find any solution to the problem?
I have gappy data for wich I want to make a svd for
estimating seasonal variability. Reconstruction of the gaps
is not possible though.
Any help is welcome.
Thanks
Heiko