diff options
Diffstat (limited to 'vignettes/gmkin_manual.Rmd')
| -rw-r--r-- | vignettes/gmkin_manual.Rmd | 114 |
1 files changed, 57 insertions, 57 deletions
diff --git a/vignettes/gmkin_manual.Rmd b/vignettes/gmkin_manual.Rmd index 3a0f3f0..45dae83 100644 --- a/vignettes/gmkin_manual.Rmd +++ b/vignettes/gmkin_manual.Rmd @@ -20,17 +20,17 @@ opts_chunk$set(tidy = FALSE, cache = FALSE) ## Introduction The R add-on package gmkin provides a browser based graphical interface for -performing kinetic evaluations of degradation data using the +performing kinetic evaluations of degradation data using the [mkin package](http://kinfit.r-forge.r-project.org/mkin_static). While the use of gmkin should be largely self-explanatory, this manual may serve as a functionality overview and reference. -For system requirements and installation instructions, please refer to the +For system requirements and installation instructions, please refer to the [gmkin homepage](http://kinfit.r-forge.r-project.org/gmkin_static). ## Starting gmkin -As gmkin is an R package, you need to start R and load the gmkin package before you can run gmkin. +As gmkin is an R package, you need to start R and load the gmkin package before you can run gmkin. The latter can be achieved by entering the command ```{r, eval = FALSE} @@ -39,7 +39,7 @@ library(gmkin) into the R console. This will also load the packages that gmkin depends on, most notably gWidgetsWWW2 and mkin. Loading the package only has to be done -once after you have started R. +once after you have started R. Before you start gmkin, you should make sure that R is using the working directory that you would like to keep your gmkin project file(s) in. If you use @@ -81,7 +81,7 @@ However, it also contains a tab 'Data' for editing kinetic data. ## Project file management -At startup, the project explorer to the left shows the two project workspaces 'FOCUS_2006' and +At startup, the project explorer to the left shows the two project workspaces 'FOCUS_2006' and 'FOCUS_2006_Z' delivered with the package. The project manangement area in the center gives the possibility to save these projects under a new name, or to start a new, empty project. @@ -90,19 +90,19 @@ A gmkin project workspace contains datasets, kinetic models for fitting, and so-called fits, i.e. the results of fitting models to data. The project area also shows the current working directory, where project workspace files are saved using the file extension `.gmkinws`. -Once a project has been saved by the user, the project explorer to the left will show it in the +Once a project has been saved by the user, the project explorer to the left will show it in the project list.  The current state of a project should repeatedly be saved during the work in order to avoid -loosing data. This can be achieved by selecting the 'Project' tab in the center and pressing +loosing data. This can be achieved by selecting the 'Project' tab in the center and pressing the 'Save project to project file' button. More conveniently, the keyboard shortcut Shift-F12 (was Ctrl-X in gmkin < 0.6.6) can be used to save the current status of the project. -In the project file management area, datasets and models can be imported from -one of the projects in the project list, once it has been selected in the +In the project file management area, datasets and models can be imported from +one of the projects in the project list, once it has been selected in the droplist labelled 'Import from:'.  @@ -114,7 +114,7 @@ if you are limited in vertical screen space. ## Dataset editor -When you select one of the datasets in the dataset explorer to the left, some +When you select one of the datasets in the dataset explorer to the left, some summary information about the dataset is shown in the center, and the data itself is loaded into the data editor to the right. @@ -123,8 +123,8 @@ is loaded into the data editor to the right. When you have added information about the units, or edited the data to the right, you should hit the button 'Keep changes'. This will add a new entry in the dataset explorer. -In the dataset editor to the right, you can override original data (for example in order to -follow FOCUS recommendations for time zero samples or values below the limit of detection) +In the dataset editor to the right, you can override original data (for example in order to +follow FOCUS recommendations for time zero samples or values below the limit of detection) by entering numbers in the override column. When saving a dataset, the sampling times, the number of replicates and the list of observed @@ -158,7 +158,7 @@ be able to use the dataset that you created in a future gmkin session. ### Importing data from text files -In case you want to work with a larger dataset that is already available as a computer +In case you want to work with a larger dataset that is already available as a computer file e.g. in a spreadsheet application, you can export these data as a tab separated or comma separated text file and import it using the "Browse" and "Upload" buttons in the dataset editor. @@ -167,14 +167,14 @@ As an example, we can create a text file from one of the datasets shipped with the mkin package using the following R command: ```{r, eval = FALSE} -write.table(schaefer07_complex_case, sep = ",", dec = ".", - row.names = FALSE, quote = FALSE, +write.table(schaefer07_complex_case, sep = ",", dec = ".", + row.names = FALSE, quote = FALSE, file = "schaefer07.csv") ``` -This produces a text file with comma separated values in the current working directory of R. +This produces a text file with comma separated values in the current working directory of R. -Loading this text file into gmkin using the "Browse" and "Upload" buttons +Loading this text file into gmkin using the "Browse" and "Upload" buttons  @@ -192,14 +192,14 @@ the column names, i.e. the names of the observed variables when the data are in As "Separator", whitespace, semicolon or comma can be chosen. If whitespace is selected, files in which the values are separated by a fixed or varying number of whitespace characters -should be read in correctly. As the tabulator counts as a whitespace character, this is +should be read in correctly. As the tabulator counts as a whitespace character, this is also the option to choose for tabulator separated values. As the "Decimal" separator, comma "," or period "." can be selected. In the next line, it can be specified if the data are in wide or in long format. If in wide format, the only option left to specify is the title of the column containing -the sampling times. If the data is in long format, the column headings specifying the +the sampling times. If the data is in long format, the column headings specifying the columns containing the observed variables (default is "name"), the sampling times (default is "time"), the observed values (default is "value") and, if present in the data, the relative errors (default is "err") can be adapted. The default settings appearing if @@ -223,7 +223,7 @@ in a future gmkin session. ## Model editor -The following screenshot shows the model editor for the model 'Z.2a.ff' from +The following screenshot shows the model editor for the model 'Z.2a.ff' from the project workspace 'FOCUS_2006_Z' provided with the gmkin package.  @@ -255,14 +255,14 @@ If a dataset and a kinetic model are selected, the button "Configure fit" below in the Explorer window "Configuration" becomes active. Pressing it opens the "Configuration" area in the center and loads the dataset into the Data editor to the right, for viewing only. -After pressing the button 'Plot unoptimised', the data and the kinetic model with default +After pressing the button 'Plot unoptimised', the data and the kinetic model with default starting parameters are plotted in the 'Plot' area to the right.  ### Fit options -The most important fit options of the `mkinfit` function can be set via the +The most important fit options of the `mkinfit` function can be set via the controls in the fit configuration area shown above. If the "plot" checkbox is checked (and on Windows, the R GUI is used for running gmkin), an R graphics device started via the R console shows the fitting progress, i.e. the change of @@ -270,7 +270,7 @@ the model solution together with the data during the optimisation. The "solution\_type" can either be "auto", which means that the most effective solution method is chosen for the model, in the order of "analytical" (for parent only degradation -data), "eigen" (for differential equation models with only linear terms, i.e. without +data), "eigen" (for differential equation models with only linear terms, i.e. without FOMC, DFOP or HS submodels) or "deSolve", which can handle all model definitions generated by the `mkin` package. If a compiler (gcc, on Windows install the 'Rtools' software package), then the "deSolve" solution method is alway chosen when there are more than one @@ -280,8 +280,8 @@ The parameters "atol" and "rtol" are only effective if the solution type is "deS control the precision of the iterative numerical solution of the differential equation model. The checkboxes "transform\_rates" and "transform\_fractions" control if the parameters are fitted -as defined in the model, or if they are internally transformed during the fitting process in -order to improve the estimation of standard errors and confidence intervals which are based +as defined in the model, or if they are internally transformed during the fitting process in +order to improve the estimation of standard errors and confidence intervals which are based on a linear approximation at the optimum found by the fitting algorithm. If fitting with transformed fractions leads to a suboptimal fit, doing a first @@ -289,9 +289,9 @@ run without transforming fractions may help. A final run using the optimised parameters from the previous run as starting values (see comment on "Get initials from" above) can then be performed with transformed fractions. -The dropdown box "weight" specifies if and how the observed values should be weighted -in the fitting process. If "manual" is chosen, the values in the "err" column of the -dataset are used, which are set to unity by default. Setting these to higher values +The dropdown box "weight" specifies if and how the observed values should be weighted +in the fitting process. If "manual" is chosen, the values in the "err" column of the +dataset are used, which are set to unity by default. Setting these to higher values gives lower weight and vice versa. If "none" is chosen, observed values are not weighted. Please refer to the documentation of the `modFit` function from the `FME` package for the meaning of options "std" and "mean". @@ -305,7 +305,7 @@ for more details. IRLS fitting can be configured using the options The drop down box "method.modFit" makes it possible to choose between the optimisation algorithms "Port" (the default in mkin versions > 0.9-33, a local optimisation algorithm using a model/trust region approach), "Marq" (the former default in -mkin, a Levenberg-Marquardt variant from the R package `minpack.lm`), +mkin, a Levenberg-Marquardt variant from the R package `minpack.lm`), and "SANN" (the simulated annealing method - robust but inefficient and without a convergence criterion). @@ -321,21 +321,21 @@ the model fit should be plotted can be selected as shown below.  On systems running the Windows operating system, the windows metafile (wmf) format -can be additionally chosen. Changing the file format for plotting will also change +can be additionally chosen. Changing the file format for plotting will also change the extension of the proposed filename for saving the plot. ### Parameters -Below the fit control area, a table with the Starting parameter list is displayed. While +Below the fit control area, a table with the Starting parameter list is displayed. While name and type of the parameters should not be edited, their initial values can be edited by clicking on a row. Also, it can be specified if the parameters should be fixed -in the optimisation process. +in the optimisation process. If the initial values for the parameters were changed, the resulting model solution -can be visually checked by pressing the button "Plot unoptimised". This will update the +can be visually checked by pressing the button "Plot unoptimised". This will update the plot of the model and the data using the specified initial parameter values. -If a similar model with a partially overlapping model definition has already be fitted, +If a similar model with a partially overlapping model definition has already be fitted, initial values for parameters with the same name in both models can also be retrieved from previous fits by selecting the fit and pressing the button "Get starting parameters from". This facilitates stepwise fitting of more complex degradation pathways. @@ -347,26 +347,26 @@ parameter values are added to the parameter table. In many cases the starting parameters and the fit options do not need to be modified and the model fitting process can simply be started by pressing the "Run fit" button. -In the R console, the progressive reduction in the model cost can be monitored and will +In the R console, the progressive reduction in the model cost can be monitored and will be displayed in the following way: ```{r, eval = FALSE} -Model cost at call 1 : 15156.12 -Model cost at call 3 : 15156.12 -Model cost at call 7 : 14220.79 -Model cost at call 8 : 14220.79 -Model cost at call 11 : 14220.79 -Model cost at call 12 : 3349.268 -Model cost at call 15 : 3349.268 -Model cost at call 17 : 788.6367 -Model cost at call 18 : 788.6366 -Model cost at call 22 : 374.0575 -Model cost at call 23 : 374.0575 -Model cost at call 27 : 371.2135 -Model cost at call 28 : 371.2135 -Model cost at call 32 : 371.2134 -Model cost at call 36 : 371.2134 -Model cost at call 37 : 371.2134 +Model cost at call 1 : 15156.12 +Model cost at call 3 : 15156.12 +Model cost at call 7 : 14220.79 +Model cost at call 8 : 14220.79 +Model cost at call 11 : 14220.79 +Model cost at call 12 : 3349.268 +Model cost at call 15 : 3349.268 +Model cost at call 17 : 788.6367 +Model cost at call 18 : 788.6366 +Model cost at call 22 : 374.0575 +Model cost at call 23 : 374.0575 +Model cost at call 27 : 371.2135 +Model cost at call 28 : 371.2135 +Model cost at call 32 : 371.2134 +Model cost at call 36 : 371.2134 +Model cost at call 37 : 371.2134 Optimisation by method Port successfully terminated. ``` @@ -374,7 +374,7 @@ If plotting of the fitting progress was selected, a new separate graphics window should either pop up, or a graphics window previously started for this purpose will be reused. -If your screen size allows for it, you can arrange the R plotting window and the +If your screen size allows for it, you can arrange the R plotting window and the R console in a way that you can see everything at the same time:  @@ -382,7 +382,7 @@ R console in a way that you can see everything at the same time: ## Results and summary Once a fit has successfully been completed, the most important results are shown in several -tables in the "Result" area as shown below. +tables in the "Result" area as shown below.  @@ -399,8 +399,8 @@ Whenever a new fit has been configured or a run of a fit has been completed, the area is updated with the abovementioned plot of the data and the current model solution. In addition, a confidence interval plot is shown below this conventional plot. In case -a fit has been run and confidence intervals were successfully calculated for the fit (i.e. -if the model was not overparameterised and no other problems occurred), the +a fit has been run and confidence intervals were successfully calculated for the fit (i.e. +if the model was not overparameterised and no other problems occurred), the confidence intervals are graphically displayed as bars as shown below.  @@ -408,8 +408,8 @@ confidence intervals are graphically displayed as bars as shown below. ## Howtos The following sections show step by step descriptions of how to perform certain -tasks using gmkin. In principle, this should be necessary as the GUI was -designed to be largely self-explanatory. Nevertheless may help a beginner to +tasks using gmkin. In principle, this should not be necessary as the GUI was +designed to be largely self-explanatory. Nevertheless it may help a beginner to understand how to use gmkin. At the same time, the gmkin author uses them as test cases to make sure that the most important functionality is not broken before releasing a new version. |