aboutsummaryrefslogtreecommitdiff
path: root/R
AgeCommit message (Collapse)AuthorFilesLines
2007-03-29Now using the "port" algorithm for the model fitting ranke1-4/+4
by nls, because in some cases it converges when the default algorithm does not. git-svn-id: http://kriemhild.uft.uni-bremen.de/svn/drfit@87 d1b72e20-2ee0-0310-a1c4-ad5adbbefcdc
2006-08-28Fixed the line types in drplot. Now we get different lines ranke1-4/+8
for different fits (same substance), and plotting for multiple substances without overlay as needed for drfit-tox works again. git-svn-id: http://kriemhild.uft.uni-bremen.de/svn/drfit@86 d1b72e20-2ee0-0310-a1c4-ad5adbbefcdc
2006-08-28Fix for the weibull fitting. Now there is a warning if ranke1-18/+26
the log ED 50 can not be found, and the log ED50 is not reported. git-svn-id: http://kriemhild.uft.uni-bremen.de/svn/drfit@85 d1b72e20-2ee0-0310-a1c4-ad5adbbefcdc
2006-08-14- New version just published on my websiteranke4-58/+109
- Fixes in checkplate, checksubstance and checkexperiment - New arguments ltys, xlab and ylab in drplot, due to a request by Ewa Mulkiewicz git-svn-id: http://kriemhild.uft.uni-bremen.de/svn/drfit@83 d1b72e20-2ee0-0310-a1c4-ad5adbbefcdc
2006-05-22Make the logit model work again, by setting a sensibleranke1-8/+9
starting value for the scale argument. Now all models get separate starting arguments, with hopefully effective default values. git-svn-id: http://kriemhild.uft.uni-bremen.de/svn/drfit@80 d1b72e20-2ee0-0310-a1c4-ad5adbbefcdc
2006-05-08- Corrected the use of the par() function, so no ranke2-9/+3
unnecessary devices are opened any more - Adjusted the .Rout.save files to kriemhilds version (R 2.3.0 from Debian unstable, i386) git-svn-id: http://kriemhild.uft.uni-bremen.de/svn/drfit@78 d1b72e20-2ee0-0310-a1c4-ad5adbbefcdc
2006-05-05- Fix of the drfit() function for the case that the calculationranke1-2/+4
of the confidence interval fails - Updated version of the drfit-Rnews preprint git-svn-id: http://kriemhild.uft.uni-bremen.de/svn/drfit@77 d1b72e20-2ee0-0310-a1c4-ad5adbbefcdc
2006-05-03Changed the way confidence intervals are being calculated. Nowranke1-42/+55
the function confint.nls from the MASS package is being used, after stumbling across the relevant paragraph of the MASS book. git-svn-id: http://kriemhild.uft.uni-bremen.de/svn/drfit@76 d1b72e20-2ee0-0310-a1c4-ad5adbbefcdc
2006-04-28I also removed the get(getOption("device"))()ranke1-2/+8
call from the checkexperiment function, and updated the INDEX. git-svn-id: http://kriemhild.uft.uni-bremen.de/svn/drfit@75 d1b72e20-2ee0-0310-a1c4-ad5adbbefcdc
2006-04-28- Fixed the drplot function, so the graphics parameter ask is adequately setranke1-3/+6
depending on the overlay argument, and set back to the previous value - Added a test for each datase - Added a demo for each dataset - Corrected many small bugs in the documentation and improved formatting of the .Rd files git-svn-id: http://kriemhild.uft.uni-bremen.de/svn/drfit@74 d1b72e20-2ee0-0310-a1c4-ad5adbbefcdc
2006-04-26- drplot: Removed the call to get(getOption("device")), since it causedranke1-11/+5
problems when calling drplot from Sweave code chunks as well as in JGR. - drplot: Removed the output with cat, since it may be annoying to receive a message when a plot was requested - Small documentation adaptations and fixes git-svn-id: http://kriemhild.uft.uni-bremen.de/svn/drfit@70 d1b72e20-2ee0-0310-a1c4-ad5adbbefcdc
2006-04-21Added the checkplate.R function, which is now just another wayranke1-0/+4
to call checkexperiment.R, and fixed DESCRIPTION and INDEX git-svn-id: http://kriemhild.uft.uni-bremen.de/svn/drfit@67 d1b72e20-2ee0-0310-a1c4-ad5adbbefcdc
2006-04-21Working version of the checkexperiment function, ranke2-113/+50
including documentation, but R CMD check tells me that "use of NULL environment is deprecated" git-svn-id: http://kriemhild.uft.uni-bremen.de/svn/drfit@66 d1b72e20-2ee0-0310-a1c4-ad5adbbefcdc
2006-04-21First chunk of work for making checkplate an aliasranke1-0/+101
to checkexperiment which will also work for the ecotox database. git-svn-id: http://kriemhild.uft.uni-bremen.de/svn/drfit@65 d1b72e20-2ee0-0310-a1c4-ad5adbbefcdc
2006-04-03- The drplot function now accepts custom xlim and ylim valuesranke2-15/+41
- Confidence intervals for the EC50 can now (again) be generated by drfit by giving the argument conf=TRUE - Update of the INDEX git-svn-id: http://kriemhild.uft.uni-bremen.de/svn/drfit@62 d1b72e20-2ee0-0310-a1c4-ad5adbbefcdc
2006-04-01Corrected the split of the drift.R file. ranke3-288/+287
Added example data IM1xIPC81. git-svn-id: http://kriemhild.uft.uni-bremen.de/svn/drfit@61 d1b72e20-2ee0-0310-a1c4-ad5adbbefcdc
2006-03-30Split of the drfit.R file to have a separate file for eachranke5-621/+619
function. git-svn-id: http://kriemhild.uft.uni-bremen.de/svn/drfit@60 d1b72e20-2ee0-0310-a1c4-ad5adbbefcdc
2006-03-29- Added an error message for drplotranke1-0/+5
- Recalculated the XY data, so the response data is now calculated based on growth rates instead of simple frond numbers - Reduced the number of columns in the IM1xVibrio data git-svn-id: http://kriemhild.uft.uni-bremen.de/svn/drfit@59 d1b72e20-2ee0-0310-a1c4-ad5adbbefcdc
2006-03-16Added the pyrithione and IM1x V. fischeri data with documentationranke1-1/+1
and cleaned the antioul data from some unnecessary components. git-svn-id: http://kriemhild.uft.uni-bremen.de/svn/drfit@58 d1b72e20-2ee0-0310-a1c4-ad5adbbefcdc
2006-03-06Added the drfit-package help page, a new sample data fileranke1-8/+23
and a preliminary version of the drfit-Rnews article. R/drfit.R - the possibility to directly write pdf graphs was added - the results dataframe now has numbered lines where it didn't The checkplate and checksubstance functions are now classified internal in their help files. git-svn-id: http://kriemhild.uft.uni-bremen.de/svn/drfit@57 d1b72e20-2ee0-0310-a1c4-ad5adbbefcdc
2006-02-27Now the model type (mtype) is set to "active" if the response at the lowestranke1-109/+118
dose level is smaller than 0.5. git-svn-id: http://kriemhild.uft.uni-bremen.de/svn/drfit@56 d1b72e20-2ee0-0310-a1c4-ad5adbbefcdc
2006-01-02A fix for a bug that kept drplot from plotting if no control data ranke1-3/+2
are present in the dose-response data frame. git-svn-id: http://kriemhild.uft.uni-bremen.de/svn/drfit@55 d1b72e20-2ee0-0310-a1c4-ad5adbbefcdc
2006-01-02- Added a little documentation of the sample data in inst/doc.ranke1-10/+36
- Added the possibility to show standard deviation or confidence interval for the mean of the control values, if present in the dose-response data. Also, control data (with dose=0) are not listed in the results data frame of drfit any more. git-svn-id: http://kriemhild.uft.uni-bremen.de/svn/drfit@54 d1b72e20-2ee0-0310-a1c4-ad5adbbefcdc
2005-12-22The calculation of dose-response curves in drfit was wrong if the number ofranke1-5/+14
dose levels was different from 9. Now, n is calculated as the rounded mean number of response data per dose level. The number of dose levels ndl is additionally reported. The test was removed because it didn't work properly. git-svn-id: http://kriemhild.uft.uni-bremen.de/svn/drfit@52 d1b72e20-2ee0-0310-a1c4-ad5adbbefcdc
2005-12-21Changed the terminology from EC50 to ED50 reflecting that the package is meantranke1-44/+44
for dose-response analysis in general and not only for concentration-response analysis. This is the first revision that has been prepared with the knowledge of the existence of the drc package for dose-response curve analysis. git-svn-id: http://kriemhild.uft.uni-bremen.de/svn/drfit@48 d1b72e20-2ee0-0310-a1c4-ad5adbbefcdc
2005-11-25Fixed the check for dose == 0 in drplot() and improvedranke1-5/+11
the model naming in the result table given by drfit() git-svn-id: http://kriemhild.uft.uni-bremen.de/svn/drfit@47 d1b72e20-2ee0-0310-a1c4-ad5adbbefcdc
2005-10-11Go back to the png device, since we get problems with "(" and ")" in theranke1-4/+4
filenames with ghostscript. git-svn-id: http://kriemhild.uft.uni-bremen.de/svn/drfit@45 d1b72e20-2ee0-0310-a1c4-ad5adbbefcdc
2005-10-11Change the device for creation of png files to bitmap(), in orderranke1-4/+4
to let people without the X11() device make pngs, too. git-svn-id: http://kriemhild.uft.uni-bremen.de/svn/drfit@43 d1b72e20-2ee0-0310-a1c4-ad5adbbefcdc
2005-10-10Percent standard deviation of controls is given by the output of theranke1-3/+4
checkplate function. git-svn-id: http://kriemhild.uft.uni-bremen.de/svn/drfit@41 d1b72e20-2ee0-0310-a1c4-ad5adbbefcdc
2005-10-10Change the position of the legend in the checkplate and checksubstance functionsranke1-2/+2
to topleft instead of topright. git-svn-id: http://kriemhild.uft.uni-bremen.de/svn/drfit@40 d1b72e20-2ee0-0310-a1c4-ad5adbbefcdc
2005-09-16Added support for the UFT ecotox database on eckehaat.ranke1-5/+13
git-svn-id: http://kriemhild.uft.uni-bremen.de/svn/drfit@36 d1b72e20-2ee0-0310-a1c4-ad5adbbefcdc
2005-08-02drplot: Added the pointsize argument for postscript and png graphsranke1-21/+22
drfit: Changed the order of fits in the case of chooseone=TRUE git-svn-id: http://kriemhild.uft.uni-bremen.de/svn/drfit@34 d1b72e20-2ee0-0310-a1c4-ad5adbbefcdc
2005-07-27Added the pointsize argument to drplot.ranke1-7/+8
git-svn-id: http://kriemhild.uft.uni-bremen.de/svn/drfit@31 d1b72e20-2ee0-0310-a1c4-ad5adbbefcdc
2005-07-18Now the weibull model has an automatic choice of starting values, fixing a ↵ranke1-5/+4
bug that prevented fitting the weibull function in most cases in the last version. Also, the log EC50 is reported as estimated with nlm from the fitted weibull model. Because of the important changes, mainly in naming the different models which leads to incompatibility with the old drfit syntax, I am increasing the minor version number to 0.04 with this release. git-svn-id: http://kriemhild.uft.uni-bremen.de/svn/drfit@30 d1b72e20-2ee0-0310-a1c4-ad5adbbefcdc
2005-07-16Several changes in this revision:ranke1-48/+87
- Addition of the weibull fit - Renaming of lognorm to probit, logis to logit, and linearlogis to linlogit - Fix of a bug that prevented the logis fit to work in recent version - Different way of listing parameters in the output. - Documentation of changes git-svn-id: http://kriemhild.uft.uni-bremen.de/svn/drfit@29 d1b72e20-2ee0-0310-a1c4-ad5adbbefcdc
2005-07-15Some cleanup of the legend positioning, which can be used starting fromranke1-14/+5
R version 2.1.0. git-svn-id: http://kriemhild.uft.uni-bremen.de/svn/drfit@28 d1b72e20-2ee0-0310-a1c4-ad5adbbefcdc
2005-07-15drplot now uses the "topright" argument to specify the position of the legend,ranke1-2/+2
both for single substance plots and for overlay plots. git-svn-id: http://kriemhild.uft.uni-bremen.de/svn/drfit@27 d1b72e20-2ee0-0310-a1c4-ad5adbbefcdc
2005-07-07Added the possibility to simultaneously fit dose-response curves with differentranke1-3/+19
units in drfit. Added cat output from drfit to keep track of which substance is currently being worked on. git-svn-id: http://kriemhild.uft.uni-bremen.de/svn/drfit@26 d1b72e20-2ee0-0310-a1c4-ad5adbbefcdc
2005-05-19Added the "allWrong" argument to the drfit function for dose-response data ↵ranke1-4/+7
that produce misleading or wrong fits with all dose-response models. git-svn-id: http://kriemhild.uft.uni-bremen.de/svn/drfit@25 d1b72e20-2ee0-0310-a1c4-ad5adbbefcdc
2005-04-26The drdata and the drfit function can now handle data with "no fit" entries inranke1-3/+4
the "ok" column. They are not used for fitting, but they are plotted. git-svn-id: http://kriemhild.uft.uni-bremen.de/svn/drfit@21 d1b72e20-2ee0-0310-a1c4-ad5adbbefcdc
2005-03-17The filename generation from substance names in drplot was changed. Now, ranke1-4/+5
parentheses () are not substituted any more. Only whitespace is substituted by "_". git-svn-id: http://kriemhild.uft.uni-bremen.de/svn/drfit@20 d1b72e20-2ee0-0310-a1c4-ad5adbbefcdc
2005-03-14I added the possibility to explicitly position or exclude the ranke1-521/+530
legend(s). git-svn-id: http://kriemhild.uft.uni-bremen.de/svn/drfit@19 d1b72e20-2ee0-0310-a1c4-ad5adbbefcdc
2005-03-14Added the option to leave the graphics file device open in drplot().ranke1-2/+6
git-svn-id: http://kriemhild.uft.uni-bremen.de/svn/drfit@18 d1b72e20-2ee0-0310-a1c4-ad5adbbefcdc
2005-02-25Some bugfixes and the addition of the linearlogisWrong argumentranke1-86/+91
to the drfit function. git-svn-id: http://kriemhild.uft.uni-bremen.de/svn/drfit@16 d1b72e20-2ee0-0310-a1c4-ad5adbbefcdc
2005-02-25rewritten drplot function, documentation not up to dateranke1-140/+100
git-svn-id: http://kriemhild.uft.uni-bremen.de/svn/drfit@14 d1b72e20-2ee0-0310-a1c4-ad5adbbefcdc
2005-02-24I added support for the enzyme test database.ranke1-35/+63
git-svn-id: http://kriemhild.uft.uni-bremen.de/svn/drfit@13 d1b72e20-2ee0-0310-a1c4-ad5adbbefcdc
2005-02-24I added the possibility to select one from the possible fits in theranke1-51/+60
drfit function, so only one fit is reported per substance. git-svn-id: http://kriemhild.uft.uni-bremen.de/svn/drfit@12 d1b72e20-2ee0-0310-a1c4-ad5adbbefcdc
2005-02-23I addded the checksubstance function, and I improved the ranke1-100/+169
drfit function, so it now also works with larger datasets, and if some substances have no data. git-svn-id: http://kriemhild.uft.uni-bremen.de/svn/drfit@11 d1b72e20-2ee0-0310-a1c4-ad5adbbefcdc
2004-11-19I improved the color defaults in drplot, so under simple circumstancesranke1-9/+15
the same colors are chosen for data and fited curves. git-svn-id: http://kriemhild.uft.uni-bremen.de/svn/drfit@9 d1b72e20-2ee0-0310-a1c4-ad5adbbefcdc
2004-09-16I added the odbcClose statement, since obviously the number of open channels(no author)1-0/+3
is restricted. JR git-svn-id: http://kriemhild.uft.uni-bremen.de/svn/drfit@7 d1b72e20-2ee0-0310-a1c4-ad5adbbefcdc

Contact - Imprint