|
| |
|
Workshop by
The
EWRS Education and Training Working Group’s
«Nonlinear regression problems in weed science with the free software R»
Time: Sunday 22nd of June 2008, 8:00-16:00
In conjunction with the forthcoming International Weed Science Congress,
Vancouver, Canada.
Click here
for details on how to apply for a reduced fee |
Course Description
Herbicides are designed to kill plants and their
selectivity in crops is a question of dose. Thus, non-linear
dose-response curves are fundamental to understanding herbicide
selectivity and efficacy. This principle can be extended to other
branches of weed management, e.g. allelopathy, flame weeding, weed
steaming, physical control, laser cutting, radiation, weed crop
competition and germination curves. The aim is to make participants
able to succinctly assess, for example, toxicity weed crop
competition on the basis of proper choice of response models and
compare response curves statistically using the free software
programme R and a specific developed extension package, drc,
developed for weed scientists and freely downloadable at URL:
www.bioassay.dk. Additionally, the package drc can also be used to
fi t germination curves and user defined functions, for example
competition models.
Course Content
 |
The concept of nonlinear response curves |
|
Installation of R and how to work with the programme.
|
|
Fitting and comparing response curves, e.g.: discrete versus
gradual endpoints, relative versus raw data, upper and lower
limits, measurement of efficacy and index of selectivity of
herbicides, competitive ability, germination ability and
bioassay design, heterogeneity of variance.
|
|
Working on handout data or, preferentially, own data.
|
Fee and Duration
Eight hours; IWSS members USD $100, others USD $150
Participants
Graduates and postgraduate students, weed scientists in public
and private institutions. The participants are strongly encouraged
to bring their own data.
There will be a
maximum of 30 attendants on a first come, first-served basis.
Course Material
Course material will be uploaded at the www.bioassay.dk website
one week prior to the workshop. Participants must bring their own
computer.
Prerequisite
Basic knowledge of statistics, including regression analysis, would
be preferable. |
|
Programme
|
Time |
Headline |
Content |
|
8:00-8:30 |
Welcome |
Welcome Introduction of teachers and participants. Overview of
short course |
|
8:30-9:15 |
Biological aspects |
Choice of endpoint, dose response models, parameters of interest |
|
9:15-10:00 |
Modeling aspects |
model assumptions, diagnostics, remedies for model deviations. |
|
10:00-10:30 |
Coffee/tea break |
|
|
10:30-11:00 |
Installation and configuration of R |
Installation of R on the participants |
|
11:00-11:45 |
Introduction to R |
The
R language: objects, basic structures and functions,
graphics, add-on packages. |
|
11:45-12:30 |
First steps in R |
A
few exercises to get started: reading data and getting help. |
|
12:30-13:30 |
Lunch |
|
|
13:30-15:00 |
Case studies |
Examples of estimation and model check, fitting multiple
dose-response curves, comparison of relevant parameters,
interpretation and visualization of results.
|
|
15:00- 15:55 |
Analysis of data |
Participants use own data or hand-out data and carry out a
complete analysis. Analyses will depend on participants’
interests and problems.
|
|
15:55 |
Conclusion |
Evaluation of the day |
* It is advised that participants, who
bring their own data, arrange data in a comma-separated file format (csv)
with one column per variable, and one row per observation. |
|
