Grokbase Groups R r-help April 2012
FAQ
Hey all,

I need some help with a cross validation. I'm new with R and as well with
statistics. I had a group work to create a tool for remote sensing class
that extracts the best bands of hyperspectral satellite images that describe
vegetation. Its a regression between a linear function of using a normalized
differenced index (i-j)/(i+j) while i and j are the bands (in the data these
are the columns, expect the last column) and the ground truth data which is
listed in the last column in %.
We did a manual cross validation (described below), but as the code is too
long and confusing, we'd like to use the cv.glm function out of the boot
package. We've tried it several times, but we don't know how to do ist.
Could anybody help us?
This is our current code for the tool with a manual cross validation:


Thanks a lot,
Motte

--
View this message in context: http://r.789695.n4.nabble.com/Package-boot-funtion-cv-glm-tp4545275p4545275.html
Sent from the R help mailing list archive at Nabble.com.

Search Discussions

  • Drew Tyre at Apr 10, 2012 at 3:37 pm
    The key to using cv.glm is that you have to have a fitted model object with
    all the data to validate. In your case, that would appear to be a model
    like this:
    lm(base[,ncol(base)]~NDI)
    where NDI is calculated from two bands in the dataframe base. However, if
    the ground truth data is independently collected from the imagery (the
    usual case), then there is no need to do cross validation - the model above
    is all you need? If you repeat the model above for all pairwise
    combinations of bands, then comparing the models using R^2 or some other
    metric would tell you which band combination is superior.

    Cheers
    On Tue, Apr 10, 2012 at 6:01 AM, Motte wrote:

    Hey all,

    I need some help with a cross validation. I'm new with R and as well with
    statistics. I had a group work to create a tool for remote sensing class
    that extracts the best bands of hyperspectral satellite images that
    describe
    vegetation. Its a regression between a linear function of using a
    normalized
    differenced index (i-j)/(i+j) while i and j are the bands (in the data
    these
    are the columns, expect the last column) and the ground truth data which is
    listed in the last column in %.
    We did a manual cross validation (described below), but as the code is too
    long and confusing, we'd like to use the cv.glm function out of the boot
    package. We've tried it several times, but we don't know how to do ist.
    Could anybody help us?
    This is our current code for the tool with a manual cross validation:


    Thanks a lot,
    Motte

    --
    View this message in context:
    http://r.789695.n4.nabble.com/Package-boot-funtion-cv-glm-tp4545275p4545275.html
    Sent from the R help mailing list archive at Nabble.com.

    ______________________________________________
    r-help@r-project.org mailing list
    https://stat.ethz.ch/mailman/listinfo/r-help
    PLEASE do read the posting guide
    http://www.R-project.org/posting-guide.html
    and provide commented, minimal, self-contained, reproducible code.


    --
    Drew Tyre

    School of Natural Resources
    University of Nebraska-Lincoln
    416 Hardin Hall, East Campus
    3310 Holdrege Street
    Lincoln, NE 68583-0974

    phone: +1 402 472 4054
    fax: +1 402 472 2946
    email: atyre2@unl.edu
    http://snr.unl.edu/tyre
    http://aminpractice.blogspot.com
    http://www.flickr.com/photos/atiretoo

Related Discussions

Discussion Navigation
viewthread | post
Discussion Overview
groupr-help @
categoriesr
postedApr 10, '12 at 11:01a
activeApr 10, '12 at 3:37p
posts2
users2
websiter-project.org
irc#r

2 users in discussion

Drew Tyre: 1 post Motte: 1 post

People

Translate

site design / logo © 2022 Grokbase