NEWS | R Documentation |
merge_RLum.Results()
clear break implemented when objects cannot be merged.
analyse_SAR.CWOSL()
Add support for ...
argument for calc_OSLLxTxRatio()
arguments
background.count.distribution
and sigmab
.
Analyse_SAR.OSLdata()
Add support for ...
argument for calc_OSLLxTxRatio()
arguments
background.count.distribution
and sigmab
similar as done for the function analyse_SAR.CWOSL()
.
calc_OSLLxTxRatio()
'LxTx.Error' could become negative, which is not meaningful by definition (corrected),
minor internal changes to allow for sigmab = NULL
, which has the same
effect as a missing sigmab
argument.
calc_FadingCorr()
Argument seed
added to allow to recreate previously determined results by
using a similar seed of the R random number generator,
argumentn.MCruns
no allows an option 'auto'
. Using this mode the error
estimation is continued until a stable error is reached,
added argument txtProgressBar
,
output table enhanced by input options and values resulting from the new arguments.
calc_Statistics()
Wrongly assigned output values corrected.
fit_CWCurve()
'component.contribution.matrix' is now returned as list, as otherwise the merging of the objects will break.
fit_LMCurve()
'component.contribution.matrix' is now returned as list, as otherwise the merging of the objects will break.
plot_GrowthCurve()
Function crashed with unclear error if the number of regenerations points was <=3), fixed,
fitting is now allowed with at leat two regenerations points.
plot_AbanicoPlot()
Argument na.exclude
changed to na.rm
,
add sample group specific colours for rugs,
summary and legend placement bugs for rotated abanico plot fixed,
labelling errors fixed,
argument frame
added to abanico plot,
adjustment of colours for polygon and bar in abanico plot,
summary keyword list extended and corrected.
plot_Histogram()
Argument na.exclude
changed to na.rm
,
summary keyword list extended and corrected.
plot_KDE()
Argument na.exclude
changed to na.rm
,
argument weights removed,
summary keyword list extended and corrected.
plot_RadialPlot()
Argument na.exclude
changed to na.rm
,
summary keyword list extended and corrected.
plot_RLum.Analysis()
Add ...
argument support legend.pos = "ouside"
and legend.col
for
argument combine = TRUE
writeR2BIN()
Function broke if information in SEQUENCE, COMMENT or USER was always NA. Corrected.
Switch from inlinedocs to roxygen2 to build the manuals (#193),
remove dependency to package 'sp'.
This is an unplanned update but it was necessary as with the last package version (0.4.2) unfortunately we introduced (again) a bug in the error calculation routine of the function plot_GrowthCurve(). Due to this the error of the resulting De became unrealistically small.
calc_OSLLxTxRatio()
Support for RLum.Data.Curve
objects implemented.
calc_MinDose()
Fixed a false argument in parallel::makeCluster()
that unintentionally required the 'snow' package.
analyse_SAR.CWOSL()
If the threshold for the rejection criteria was reached due an error this information was not properly stored in the output object (i.e. every time 'RC.Status' was set to 'OK' instead of 'FAILED'); corrected (#180).
extract_IrradiationTimes()
Argument file.XSYG
replaced by object
for the sake of argument consistency on the package,
allow input for RLum.Analysis
object. In such cases the arguments file.BINX
and recordType
are ignored. This mode allows for more flexibility in applying this function,
new argument compatibility.mode
to trigger the binary write output of the BIN/BINX-file,
additionally now the time between the steps is exported (TIMESINCELAST.STEP
).
fit_CWVCurve()
The function broke with an error under rare circumstances when no fit was made (reported by Alexandra Rowell); corrected,
change start parameter estimation to better account for the input data, this hopefully further stabilise the fitting (thanks to Alexandra Rowell),
more proper re-scaling of the graphical output if no fit could be produced,
new argument add fit.method
to allow using the function nlsLM
from the
minpack.lm package
and with this the Levenberg-Marquardt algorithm,
argument cex.global
replaced by cex
and moved to the ...
argument,
argument output.plot
changed to plot
,
changing the way how warnings are surpressed.
plot_Histogram()
Argument 'breaks' was without function due to a coding error. Corrected (#217).,
argument 'breaks' removed from the argument list, however, it remains supported via the '...'
argument.
plot_GrowthCurve()
Error estimation using the Monte Carlo runs for the
fitting function fit.method = "EXP+LIN"
and fit.method = "EXP+EXP"
results in unrealistic values, as the Monte Carlo simulation
was fed by mistake with wrong input variables (introduced in version 0.4.2 with the uniroot function) ... fixed,
the function broke if the uniroot algorithm was not be able to find a solution within the given ranges, fixed. Now the De will become NA insteadly and the function will not stop,
error messages coming from the fitting process itself are now surpressed.
plot_RLum.Data.Spectrum()
Colour handing changed towards a more logical behaviour for plot.type = 'single'
.
The colour now changes only the colour of the points and the argument rug
is now recognised also
by this plot type. Automatic wavelength colouring for points for in this plot type has been removed.
same changes are made for plot.type = 'multiple.lines'
,
instead of just plotting a new x-axis when the argument xaxis.energy = TRUE
is used,
the spectrum (intensity) is now proper re-calculated (#192),
minor code polishing.
plot_RLum.Anaysis()
Setting the argument combine = TRUE
for 'OSL' and 'IRSL' caused an error related
to the missing argument 'curve.transformation'; corrected.,
colour handling for setting combine = TRUE
improved. It now works as for plot.default
,
support for xlim
and ylim
in the non-combined mode,
combine mode (combine = TRUE
) is no more limited to an input object of similar curves. Now
the plot function tries to combine curves of similar type automatically.
readBIN2R()
Support for BIN/BINX-file version 07 added thanks to the support by Torben Lapp (#215).
readXSYG2R()
Info objects from the sequence header like 'name' and 'position' are now be additionally transferred to each single output object. With this the information is kept within the curve object.
writeR2BIN()
Support for BIN/BINX-file version 07 added thanks to the support by Torben Lapp (#215),
new argument compatibility.mode
to avoid position numbers >48 as this caused
errors using external software like the Analyst. The old version number is kept and added
in the comment field.
Description on the class Risoe.BINfileData
updated in accordance with the implemented
changes on the import (no change in the S4-object class chart),
added dependency to package minpack.lm
to support the Levenberg-Marquardt algorithm
for fitting in the function fit_LMCurve()
calc_IEU()
: Apply the internal-external-uncertainty (IEU) model after Thomsen et al. (2007)
to a given De distribution
extract_IrradiationTimes()
: Extract irradiation times from an XSYG file
RLum.Analysis-class
:
Fix: Argument keep.object == TRUE
was not respected for every case. After using
the function get_RLum.Analysis()
the object just contained a single object. In such
cases only the object itself, but no RLum.Analysis
object was returned (#200).
RisoeBINfileData-class
:
Slot .RESERVED
added to allow transferring undocumented raw binary values to a
new BIN/BINX-file. For this all reserved bytes are just copied and pasted into the new slot
.RESERVED
using the function readBIN2R()
. Using the function writeR2BIN()
transfers
the values unchanged to the new BIN/BINX-file. Allc corresponding functions have been updated in accordance
(readBIN2R()
, writeR2BIN()
and merge_Risoe.BINfileData()
. Other functions are not
affected as they can simply ignore the new slot.
analyse_IRSAR.RF()
Function now checks the number of channels in each curve. If the number of the channels in the first curve is > than in the 2nd curve the execution of the function is stopped,
add error estimation using bootstraping and MC loops for method = "SLIDE"
,
add scaled KDE for distribution of De,
the functions no longer stops with an error if a undefined method
was chosen. Instead
a warning message is shown and the available curves are shown in a plot anyway,
the labelling for the residuals is replaced by 'E' to avoid overplotting,
new argument rejection.criteria
provided to allow for setting rejection criteria. The
argument works similar as for other analyse_...
functions. Details are provided in the manual,
speed up caculation time for method = "SLIDE"
by re-writing the sliding process
as C++ function. It does not the affect readability of the R-code at all, as just the sum of the residual
squares is calculated in this way, i.e. the calculations are simple, but now much faster.
analyse_SAR.CWOSL()
For the case that no recycling ratio and no recuperation rate could be calculated the functions chrashed. This has bee corrected.
analyse_pIRIRSequence()
For the case that no recycling ratio and no recuperation rate could be calculated the functions chrashed. This has bee corrected; in accordance with the similar behaviour in the function analyse_SAR.CWOSL().
apply_CosmicRayRemoval()
Argument method.Pych.histogram.plot
replaced by the more general and consistent
argument plot
,
minor code polishing.
calc_FadingCorr()
Change the way the equation is solved using the R function uniroot
. This markedly
speeds up the computation process,
now additionally an RLum.Results
object is returned,
minor code polishing.
calc_MinDose()
Major changes in terms of reliablity and perfomance to the optimisation of the log likelihood function. Amongst minor fixes these include:
Optimiser of the log likelihood function changed from 'optim' to 'nlminb' and switch from 'L-BFGS-B' method to default PORT routine,
Removed parameter: boundaries
Boundaries for the model parameters are now hard-coded and suffice all input data,
Warning messages are now suppressed unless debug=TRUE
,
Improved error handling and fallback for profiling the log likelihood function,
Start value for gamma and mu are changed to lower quartile of the dose values,
fixed an error in the documentation that prevented the 'value' section from being displayed
additional information on bootstrapping added in 'details' section,
new example added,
add multicore support for bootstrapping,
new arguments: multicore
(logical), cores
(numeric).
calc_Statistics()
argument na.rm
added (#198).
CW2pHMi()
Functions was rather inflexible selecting the allowed curves, e.g. "OSL"
worked,
"OSL (UVVIS)"
not. Corrected.
CW2pLMi()
Functions was rather inflexible selecting the allowed curves, e.g. "OSL"
worked,
"OSL (UVVIS)"
not. Corrected.
CW2pLM()
Functions was rather inflexible selecting the allowed curves, e.g. "OSL"
worked,
"OSL (UVVIS)"
not. Corrected.
CW2pPMi()
Functions was rather inflexible selecting the allowed curves, e.g. "OSL"
worked,
"OSL (UVVIS)"
not. Corrected.
merge_Risoe.BINfileData()
Support for the changes in the corresponding S4-class are implemented.
plot_AbanicoPlot()
Missing summary keywords "kdemax"
, "skewness"
and "kurtosis"
added
(#205)
plot_Histogram()
Missing summary keywords "kdemax"
, "skewness"
and "kurtosis"
added
(#205)
plot_KDE()
Missing summary keywords "kdemax"
, "skewness"
and "kurtosis"
added
(#205)
plot_GrowthCurve()
Fix an error that causes a function chrash if the input data.frame (argument sample
) comprises
row names. Now all pre-set row names are removed automatically,
using the option fit.weights = FALSE
in combination with fit.method = "LIN"
caused an error. Corrected,
avoid confusing warning messages in case fit.weights
was set to FALSE
,
fix an error that causes a chrash for the combination fit.method = "LIN"
and error column just comprises 0,
argument main
was not working as expected, corrected. Furthermore the argument is removed from the
argument list, but is still available via ...
,
the slow self-written approach to calculate the De for fit.method = "EXP+LIN"
and
fit.method = "EXP+EXP"
has been replaced by the function uniroot
. This markedly reduces
the computation times for the MC simulation (partly addresses issue #100)
plot_RadialPlot()
Missing summary keywords "kdemax"
, "skewness"
and "kurtosis"
added
(#205)
plot_RLum.Analysis()
Fix an error appeared for combine = TRUE
if only one single object was in the RLum.Analysis
object (#199),
added logical argument plot.single
. In fact this just resets ncols
and nrows
to 1
(#201),
added argument curve.transformation
to allow a quick data overview using the pseudo-LM curve transformation methods.
plot_RLum.Results()
Improved error handling for RLum.Results
objects originating from calc_MinDose()
plot_RLum.Data.Spectrum()
For plot.type="multiple.lines"
the legend was placed odd, corrected using par()$usr
parameters,
argument legend.text
added for plot types which support a legend (e.g., plot.type="multiple.lines"
).
readBIN2R()
Improved handling of odd time formats,
support for the changes in the corresponding S4-class implemented.
readXSYG2R()
For using the argument recalculate.TL.curves = FALSE
the import of spectrometer
data was broken. Fixed.
during TL-curve recalculation for spectra in rare cases the produced object was not valid for further plotting due to duplicated x-values in the matrix. This has been fixed by adding artificially 1 K to such values. In addition a warnings is produced.
writeR2BIN()
Part of the manual was missing due to an error in the inlinedoc documentation. Fixed.,
the argument version
no expects a character
as import instead of a raw
object,
further suppressing of confusing warning messages,
date has not been written properly in accordance with the format description (fixed),
time values have been not properly exported in same cases, corrected.
added support for changes in corresponding S4-class, means that reserved values are transfered,
the version conversion was partly broken due to the new routines implemented with last update (fixed),
if empty comments or sample names are provided these are replaced by " "
(double empty space),
if the BIN-file is stored in a different version, the version number and the record lengths are now changed accordingly.
ExampleData.DeValues is now a list with two elements containing an additional single grain quartz data set (#203)
Adding dependency for package 'Rcpp'. This is used in particular to export some time consuming
calculations to compiled C++ code to speed up the functions. Currently it is used e.g. in the function
analyse_IRSAR.RF()
,
adding dependency (import) for the packages parallel
. With this
the multi-core support is realised within in the package. First implemented in the
function calc_MinDose()
.
merge_RLum.Data.Curve()
: Function allows merging of RLum.Data.Curve
objects in different ways (#128).
RLum.Analysis-class
:
Show warnings messages when empty object are produced by a get_RLum.Analysis()
request (implemented in get method and show method).
analyse_SAR.CWOSL()
rename argument output.plot
to plot
,
rename argument output.plot.single
to plot.single
,
rejection criteria added 'De > max. dose point'. These avoids that values are accepted even the De from the fitting should be considered as invalid.
analyse_IRSAR.RF()
Add enhanced support for ...
, e.g. log
, xlim
, ylim
,
rename argument output.plot
to plot
.
analyse_pIRIRSequence()
Fix an error that causes problems during plotting (Error in plot.new(): figure margins too large),
rename argument output.plot
to plot
,
rename argument output.plot.single
to plot.single
.
merge_RLum()
: Function extended accordingly with the newly implemented function
merge_RLum.Data.Curve()
(#128).
plot_AbanicoPlot()
:
ticks removed from error bars,
line value output removed,
placement of y- and z-labels set to center of plot,
weights-option for KDE integrated,
warning for linear z-scale removed,
fixed an error that occured in Abanico Plot for a large data range (#188),
function output extended by data points in/out of confidence interval,
histogram option added,
dot plot option added,
KDE bandwidth added to plot output,
KDE bandwidth default changed to "SJ",
rugs coloured according to subsets,
examples section extended.
plot_GrowthCurve()
:
Thanks to a querry from Rachel Smedly (#194) the De error calculation method has been updated. So far only the Lx/Tx values have been re-sampled using a Monte Carlo simlation approach to estimate the error. In contrast to the Analyst the Ln/Tn value was just considered without any error. This has been changed for all fitting methods and the final De error is calculated now also using the Ln/Tn error. Note: Usually this change results in a larger De error, however this approach is much more realistic,
consequently to the changes made in the error calculation the corresponding De resulting from
the MC simulation is now added to the output object ($De$De.MC
),
Ln/Tn point including the error is now added on the plot output as red circle,
fix a bug occasionally occurred for fit.method = "EXP+EXP"
and the function broke
subsequently (#159).
readBIN2R()
:
critical bug fixed: GRAIN
and GRAINNUMBER
were not imported properly
(thanks to Norber Mercier),
import speed has been improved by at least an order of magnitude, due to a different
internal data handling (using of data.table
instead of data.frame
). This may become
relevant for huge BIN-files like they are produced by single grain measurements (#191).
Note: The Risoe.BINfileData-class
remained untouched.
writeR2BIN()
:
minor changes due to the changes in the function readBIN2R()
,
confusing warnings during writing are suppressed.
Add new dependency for package data.table
(used in the function readBIN2R()
).
Amongst a few minor changes with this release we continued the homogenisation process to further unify our function arguments. So far meaningful similar function arguments can be used over all functions in the package to produce similar effects (e.g., na.rm = TRUE/FALSE). For you as package user it means that self-written R scripts using package functions are likely to break. Please check the list of changes below.
We apologise for any inconvenience!
RLum.Data.Image
: S4-class for handling image data including. Image
data in slot data
are stored as RasterBrick
RLum.Results
: homogenised structure for containing the following base items in the
slot data: 'summary', 'args' and 'call'. Where applicable the 'data' slot contains the original input data.
Depending on the function various further slots are added containing additional (intermediate) data and objects,
set_RLum.Data.Image()
: Set method for class RLum.Data.Image
get_RLum.Data.Image()
: get method for class RLum.Data.Image
further: corresponding show and coercing methods added for class RLum.Data.Image
.
analyse_pIRIRSequence()
: Function to analyse post-IR IRSL sequence
measurements (#133),
calc_MinDose()
: the previous calc_MinDose3()
and calc_MinDose4()
functions were merged to the new calc_MinDose()
function, which was rewritten from
scratch and now contains both the 3- and 4-parametric
Minimum Age Models (introducing the new argument par
).
Detailed changes and bugfixes are listed further below,
calc_MaxDose()
: a wrapper function that calls calc_MinDose()
,
but transforms the data set similar to the method described in Olley et al. (2006).
Rather than estimating the minimum dose of a De distribution, this function estimates the maximum dose.
calc_SourceDoseRate()
: calculating the dose rate of the irradiation source at the
date of measurement based on: source calibration date, source dose rate, dose rate error.
get_Layout()
: Helper function to return a list with layout definitions for
more homogeneous plotting,
get_RLum()
: Wrapper function for various get methods for RLum
-objects (#148),
merge_RLum.Analysis()
: Function allows merging of RLum.Analysis objects and adding
of objects an RLum.Analysis object consists of (#138),
merge_RLum()
: Wrapper function for various merge methods for RLum
-objects (#139),
plot_RLum.Data.Image()
: Plot function for RLum.Data.Image
object.
Function also added to wrapper function plot_RLum
,
plot_RLum.Results()
: Plot function for RLum.Results
objects.
RLum.Results objects created, e.g. by the age model functions may now be passed to plot_RLum.Results()
to create specific plots that previously were generated by the functions themselves.
Function also added to wrapper function plot_RLum
,
readSPE2R()
: Add support Princenton Instruments *.SPE files
(camera and spectrometer measurements). Function output is a RLum.Data.Spectrum
or RLum.Data.Image
object.
ExampleData.RLum.Data.Image
: Example data for RLum.Data.Image
object
(part of #135).
get_RLum.Analysis()
: Add support for inverse record.id
selection, e.g. record.id = -1
selects
all objects except the first.
get_RLum.Results()
:
Add support for data objects for objects
produced by the function calc_OSLLxTxRatio()
,
match access behavoir of objects produced by the function
calc_TLLxTxRatio()
. Now it works similar to the function calc_OSLLxTxRatio()
.
Risoe.BINfileData-class
: Description updated due to established changes
made for a better file format support.
Risoe.BINfileData2RLum.Analysis()
: If the pos
value
is invalid now the function does not return an error any longer. Instead
NA
is returned.
analyse_IRSAR.RF()
:
Add ...
support for mtext
,
argument method
added to allow selection between the method "FIT"
and
"SLIDE"
. The latter method has been newly implemented (#143).
analyse_SAR.CWOSL()
:
Allow 'IRSL' curves as input to support rough data analysis of the post-IR IRSL protocol (#132, Alice Versendaal, #140, Georgina King),
support multiple recuperation measurements,
correct wrong TL curve figure scaling,
argument log
moved to the ...
argument,
correct wrong regeneration point numbering (#144),
arguments cex
, main
now supported via the ...
argument,
argument mtext.outer
added,
output object Formula
added to return the formula of the
growth curve,
output object De.values
extended for D01, D02 and information
on the applied fitting,
numerical output extended for information on the choosen signal and background integration limits (#122),
enhance argument output.plot.single
for numerical values.
This allows to select each single plot output individually,
details from the argument rejection.criteria
were removed
from the argument list to allow a more flexible handling (e.g. changing only
one parameter at time) and to clean up the argument list. However, the argument
work still as expected (#145),
reformat of rejection criteria output to allow easier further treatment,
support graphical output of rejection criteria and IRSL curve (if exists),
minor bug fixes (e.g., not properly caluculated TL curve resolution) and graphical adjustements.
Analyse_SAR.OSLdata()
:
Error message implemented for non OSL data sets (#130, Georgina King),
add argument dtype
to provide further curve limitation options,
add argument keep.SEL
to allow a manual curve selection (advanced users),
minor code polishing.
calc_AliquotSize()
:
new argument plot
(logical): option to (dis-)allow plotting of the Monte Carlo estimation,
argument MC.estimate
was renamed to MC
,
new structure of the returned RLum.Results
object, containing the elements
'summary', 'args', 'call' and 'MC'.
plotting now done by plot_RLum.Results()
.
calc_CentralDose()
:
removed argument sample.id
,
removed argument print.iterations
,
renamed argument input.data
to data
,
new argument '...'
, which is checked for arguments verbose
(logical) and trace
(logical) to (dis-)allow (extended) console output,
new structure of the returned RLum.Results
object, containing the elements 'summary', 'data', 'args', 'call' and 'profile',
fixed a bug that occured for non-dispersed data sets and resulted in a virtually infinite loop using up all available memory,
plotting now done by
plot_RLum.Results().
calc_CommonDose()
:
removed argument sample.id
,
renamed argument input.data
to data
,
new argument '...'
(currently not used),
new structure of the returned RLum.Results
object, containing the elements 'summary', 'data', 'args' and 'call'.
calc_CosmicDoseRate()
:
new structure of the returned RLum.Results object, containing the elements 'summary', 'args' and 'call'
calc_FiniteMixture()
:
removed argument sample.id
,
removed argument n.iterations
,
removed argument main
,
renamed argument input.data
to data
,
new argument plot
option to (dis-)allow plotting when a vector of length > 1 is provided for n.components
,
new argument '...'
, which is checked for arguments verbose
(logical) and trace
(logical) to (dis-)allow (extended) console output. main
may be used to set the title of the plot,
new structure of the returned RLum.Results
object, containing the elements
'summary', 'data', 'args', 'call', 'mle', 'BIC', 'llik', 'grain.probability', 'components' and 'single.comp',
plotting now done by plot_RLum.Results()
.
calc_FuchsLang2001()
:
removed argument sample.mtext
,
removed argument sample.id
,
removed argument output.terminal
,
renamed argument sample
to data
,
renamed argument output.plot
to plot
,
the argument '...'
is checked for verbose
(logical) to (dis-)allow console output,
new structure of the returned RLum.Results
object, containing the elements 'summary', 'data', 'args', 'call' and 'usedDeValues',
plotting now done by plot_RLum.Results()
.
calc_HomogeneityTest()
:
removed argument sample.id
,
renamed argument input.data
to data
,
the argument '...'
is checked for verbose
(logical) to (dis-)allow console output,
new structure of the returned RLum.Results
object, containing the elements 'summary', 'data', 'args' and 'call'.
calc_MaxDose3()
:
Function renamed to calc_MaxDose()
,
this is now just a wrapper function that calls the new calc_MinDose()
function,
but transforms the data set similar to the method described in Olley et al. (2006).
For further details please see the changes/bugfixes for the calc_MinDose3()
/ calc_MinDose4()
functions,
plotting now done by plot_RLum.Results()
.
calc_MinDose3()
and calc_MinDose4()
the functions calc_MinDose3()
and calc_MinDose4()
were merged in the new function calc_MinDose()
,
renamed argument input.data
to data
,
renamed argument output.plot
to plot
,
removed arguments sample.id
, gamma.xlb
, gamma.xub
, sigma.xlb
, sigma.xub
,
mu.xlb
, mu.xub
, init.gamma
, init.sigma
, init.p0
, init.mu
, ignore.NA
,
calc.ProfileLikelihoods
, console.ProfileLikelihoods
, console.extendedOutput
, output.indices
,
new argument par
that accepts a numeric value (either 3 or 4) to set whether the 3- or 4-parametric Minimum Age Model should be applied,
new argument bootstrap
(logical) to apply the recycled bootstrap approach of Cunningham & Wallinga (2012),
new argument boundaries
that accepts a named list containing values for the lower and upper boundaries for all parameters,
new argument init.values
that accepts a named list containing the starting values for all parameters,
new argument '...'
that, amongst others, checks for debug
(logical) and verbose
(logical) to (dis-)allow (extended) console
output. Further allowed arguments are bs.M
, bs.N
, bs.h
and sigmab.sd
to set various bootstrap parameters,
implemented (recycled) bootstrap method as described in Cunningham & Wallinga (2012),
function now iteratively adjusts boundaries for sigma when the estimator reaches the upper boundary,
function now tries to estimate reasonable starting values for all parameters,
calculation and plotting of the maximum log likelihood functions/profiles are now done using the mle2() function of the 'bbmle' package
the optimisation routine was changed from 'nlminb' to 'optim' using the 'L-BFGS-B' method,
confidence intervals for all parameters are now calculated by the 'confint()' method and exclusively on the 95% confidence level
the asymmetric error on gamma (i.e., the minimum dose) is now exclusively determined on the 95% confidence interval of the profile log likelihood function for gamma
a symmetric standard error for gamma is automatically derived from the 95% confidence interval of the profile log likelihood function by dividing the length of this interval by 3.92 (see Galbraith & Roberts, 2012).
new structure of the returned RLum.Results
object, containing the elements 'summary', 'data',
'args', 'call', 'mle', 'BIC', 'confint', 'profile', 'bootstrap',
plotting now done by plot_RLum.Results()
.
calc_OSLLxTxRatio()
:
Correct an error that causes NA
values
within the curve errors for rare combinations of signal and background
integration limits,
provide a name for the output data frame LxTx.table
for the
RLum.Results
object,
error fixed for the marginal case that k = 2 and exceed the maximal curve length. In that case accidentially no error was calculated,
argument sigmab
was listed but was so far not implement. Fixed (#166)
calc_OSLLxTxRatio()
:
support of data without given uncertainties.
calc_Statistics()
:
support of data with NA
or zero errors
calc_TLLxTxRatio()
:
Lx/Tx ratio wrongly calculated as due to a coding error just the first value of the signal integral was taken for summing up the values and not the range of the values,
Output produced by the function now is a list
(data slot in the RLum object), to
equal behaviour of the functions.
fit_CWCurve()
:
Implement ...
argument to replace explicit arguments ylab
,
xlab
and main
,
Fix an issue that causes an error when input values were <= 0 (#123 reported by Johanna Lomax),
Correct wrong y-axis labelling (wrong channel resolution value).
plot_AbanicoPlot()
:
numeric values allowed for parameter centrality,
KDE label value fixed,
layout functionality added,
not supported summary keywords fixed,
dispersion polygon bug fixed,
dispersion type enriched by userdefined confidence interval,
rug option added,
multiple 2-sigma-bars allowed,
rotate-Option added to Abanico Plot,
minor modifications of layout definitions (#124),
grid.col
bug fixed (#149),
error message for data with zero or without errors added (#154),
age models as input for lines supported (#107),
support of par(mfcol())
l,
bug in lower z-axis limit eliminated,
example list updated (i.e. fast CAM instead of slow MAM)
bug for small values fixed, rounding to 0 decimals replaced by 3 significant digits,
parameter at
to allow user-defined z-tick positions,
individual statistical summary order supported (#184).
plot_DRTResults()
:
Argument na.exclude
renamed to na.rm
,
by default ylim scaling now is set to the minimum and maximum extend of the data set,
argument given.dose
now allows a vector as input, e.g., if more then one data set
is provided (#156),
argument given.dose
are not required any longer (became an optional argument).
With this the plot function can be easily re-used, e.g. for preheat plateau tests,
change behaviour of the argument na.rm
. If a data set is provided containing
NA
values also the corresponding values in provided by the argument preheat
are
cleaned up. This allows a proper usage of the function in combination with grouping option
(e.g., boxplot = TRUE
,
add argument par.local
to allow to combine output plots in broader plot
frameworks,
minor corrections in error handling and graphical appearance.
plot_Histogram()
:
Change default y-axis labelling from 'Counts' to 'Frequency',
support of par(mfcol())
(#146),
support of data without and NA
-values for error values (#152, #162),
implementation of calc_Stats()
-call for extended statistic summary,
support of summary keywords "kurtosis" and "skewness",
example y-lim bug removed (#177),
individual statistical summary order supported (#184).
plot_KDE()
:
wrong x-axis and y-axis limit calculation fixed (#158),
renaming and extension of summary and summary.pos
parameter,
plot now longer shows a density error when no error exists (#160),
De.stats error fixed (#163),
support of par(mfcol())
,
support of error-free data extended (now also for calculation of kde max) (#178),
individual statistical summary order supported (#184).
plot_GrowthCurve()
:
Minor graphical and code polishing,
fix wrong regneration point plotting, if sequence contains more then one zero dose point,
argument na.exclude
for the sake consistency replaced by na.rm
,
proper error message provided if nothing is left after NA
value removal,
overshoot of normal distribution curve in MC simulation plot corrected (#111),
argument output.plotExtended.single
added allowing a single
window plot output of all plots,
formula of the fitted function is now returend as R expression
instead of a formula
object. The return object remains, however,
Formula
,
correct wrong error message and graphical output for cases where no De can be calculated for Ln/Tn values much higher than the highest Lx/Tx value,
fix 'EXP+EXP'
fit as is was broken, due to a coding error,
correct for wrong par()
resetting,
minor code polishing and adjustment.
plot_RadialPlot()
:
support of user-defined centrality value,
rug option added,
bug for small values fixed, rounding to 0 decimals replaced by 3 significant digits,
individual statistical summary order supported (#184),
minor fixes for the graphical output.
plot_RLum.Analysis()
:
Add argument abline
to allow inserting similar vertical or horizontal lines
in every plot. The decision to implement this argument in this function instead the function
plot_RLum.Data.Curve()
was made because in a single plot modus abline
is more
sufficiently used as separate function call,
now reset par
parameters sufficiently,
add argument combine
to allow fast combining of curves in single plot window.
plot_RLum.Data.Curve()
:
Add argument norm
to allow for curve normalisation to
the highest value (#129),
Correct for insufficient information on the channel resolution when data was imported from an XSYG file.
plot_RLum.Data.Spectrum()
:
Correct an error that caused a chrash when xlim
was set
in a way that only one wavelength (colour value) was remained,
correct zebra-colour-pattern for rare cases of xlim
settings
using plot.type = "persp"
or plot.type = "persp3d"
,
add argument new plot type "transect"
for plotting a transect
along the time/temperature axis for a specific wavelength/channel interval
(#120, requested by Regina DeWitt),
correct function break for a single column matrix (single frame spectrum)
by automatically resetting plot.type
to single
for single frame spectra,
colour rug (rug = TRUE
) implemented for plot type "multiple.lines"
(#130),
argument xaxis.energy
for wavelength energy conversion for
2D plots added (#141).
readBIN2R()
:
Integrity checks added to avoid unspecific errors for wrong input formats,
complete format support as now a more throughtful distinction is made between
version 03 and 04 (also made for function writeR2BIN()
) (#182),
TAG
bit is now used to trigger the SEL
element during import.
With this, previous curve selections made with other software are kept
(#181, Norbert Mercier).
readXSYG2R()
:
If the stimulator is infrared light the recordType
will be set to IRSL
. This
makes the curve types more consistent with the commonly applied nomenclature,
now the function automatically converts time values to temperature values also
for spectra data (#121); behavious is triggered with the argument recalculate.TL.curves
.
Second2Gray()
:
Minor code and example adjustments in along with creating a new function
calc_SourceDoseRate()
(#56).
writeBIN2R()
:
Complete format support as now a more throughtful distinction is made between
version 03 and 04 (also made for function readBIN2R()
) (#182),
SEL
will now be kept using the TAG
bit, related to changes made
in function readBIN2R
(#181, Norbert Mercier).
Add new dependency for package raster
and sp
, which is used
for the RLum.Data.Image-class
,
dependency to package 'Rserve' removed,
dependcy to package 'zoo' added (used for running means),
dependency to package 'bbmle' added (used for maximum likelihood estimations).