diff --git a/DESCRIPTION b/DESCRIPTION index 719117e..ea58aa7 100644 --- a/DESCRIPTION +++ b/DESCRIPTION @@ -1,14 +1,15 @@ Package: ODEsensitivity -Version: 1.1.0 +Version: 1.1.1 Title: Sensitivity Analysis of Ordinary Differential Equations Description: Performs sensitivity analysis in ordinary differential equation (ode) models. The package utilize the ode interface from 'deSolve' and connects it with the sensitivity analysis from 'sensitivity'. Additionally we add a method to run the sensitivity analysis on variables with class 'ODEnetwork'. A detailed plotting function provides outputs on the calculations. + The method is described by Weber, Theers, Surmann, Ligges, and Weihs (2018) . Authors@R: c( - person("Stefan", "Theers", email = "stefan.theers@tu-dortmund.de", role = "aut"), person("Frank", "Weber", email = "frank.weber@tu-dortmund.de", role = "aut"), + person("Stefan", "Theers", email = "stefan.theers@tu-dortmund.de", role = "aut"), person("Dirk", "Surmann", email = "surmann@statistik.tu-dortmund.de", role = c("ctb", "cre") , comment = c(ORCID = "0000-0003-0873-137X")) ) diff --git a/NEWS.md b/NEWS.md index 6afbfe9..2fa8dcb 100644 --- a/NEWS.md +++ b/NEWS.md @@ -1,3 +1,6 @@ +# ODEsensitivity 1.1.1 +* add reference to TechReport + # ODEsensitivity 1.1.0 * add vignette * release to CRAN diff --git a/man/ODEmorris.default.Rd b/man/ODEmorris.default.Rd index 2cc85ed..12f476b 100644 --- a/man/ODEmorris.default.Rd +++ b/man/ODEmorris.default.Rd @@ -69,7 +69,7 @@ number of processor cores to be used for parallelization. Only applies if \value{ List of class \code{ODEmorris} of length \code{length(state_init)} containing in each element a matrix for one state variable. The - matrices itself contain the Morris screening results for all timepoints + matrices themselves contain the Morris screening results for all timepoints (rows: \code{mu, mu.star} and \code{sigma} for every parameter; columns: timepoints). } @@ -84,7 +84,7 @@ Function \code{\link[deSolve]{ode}} from \code{\link[deSolve]{deSolve}} is The sensitivity analysis is done for all state variables and all timepoints simultaneously using \code{\link[sensitivity]{morris}} from the - package \code{sensitivity}. + package \code{\link[sensitivity]{sensitivity}}. For non-ODE models, values for \code{r} are typically between 10 and 50. However, much higher values are recommended for ODE models (the default is diff --git a/man/ODEsobol.default.Rd b/man/ODEsobol.default.Rd index e93d0c0..a611fc5 100644 --- a/man/ODEsobol.default.Rd +++ b/man/ODEsobol.default.Rd @@ -85,10 +85,11 @@ Function \code{\link[deSolve]{ode}} from \code{\link[deSolve]{deSolve}} is The sensitivity analysis is done for all state variables and all timepoints simultaneously. If \code{sobol_method = "Jansen"}, - \code{\link[sensitivity]{soboljansen}} from the package \code{sensitivity} + \code{\link[sensitivity]{soboljansen}} from the package + \code{\link[sensitivity]{sensitivity}} is used to estimate the Sobol' sensitivity indices and if \code{sobol_method = "Martinez"}, \code{\link[sensitivity]{sobolmartinez}} - is used (also from the package \code{sensitivity}). + is used (also from the package \code{\link[sensitivity]{sensitivity}}). } \note{ If the evaluation of the model function takes too long, it might be diff --git a/man/tdcc.Rd b/man/tdcc.Rd index 509cfc7..7e727db 100644 --- a/man/tdcc.Rd +++ b/man/tdcc.Rd @@ -9,41 +9,45 @@ tdcc(ranks, pearson = FALSE, plot = FALSE) \arguments{ \item{ranks}{[\code{matrix(nrow = b, ncol = k)}]\cr \code{(bxk)}-matrix of the ranks of the \code{k} -variables for each of the \code{b} SAs, ties are neglected, +variables for each of the \code{b} sensitivity analyses, ties are neglected, must be integers.} \item{pearson}{[\code{logical(1)}]\cr -logical, should the ordinary Pearson coefficient with -Savage scores be computed (\code{b = 2})?} +Should the ordinary Pearson coefficient with +Savage scores be computed (\code{b = 2})? Default is \code{FALSE}.} \item{plot}{[\code{logical(1)}]\cr -logical, scatter plots showing rankings and Savage scores -(\code{b = 2})?} +Should scatter plots showing rankings and Savage scores be created +(\code{b = 2})? Default is \code{FALSE}.} } \value{ A named vector with components: \itemize{ - \item{\bold{kendall}}: Coefficient of concordance. - \item{\bold{pearson}}: Pearson Coefficient (only if \code{pearson = TRUE}). + \item{\code{kendall}}: Coefficient of concordance. + \item{\code{pearson}}: Pearson coefficient (only if \code{pearson = TRUE}). } } \description{ -The Top-Down Correlation Coefficient TDCC compares \code{b} rankings -using Savage Scores. +With the use of Savage scores, the Top-Down Correlation Coefficient TDCC +compares \code{b} rankings. } \details{ NOTE: As the implementation of the coefficient of concordance is still defective, please use the Pearson coefficient! } \examples{ -ranking <- rbind(1:20, - c(1,3,2,4,16,10,19,12,18,17, - 20,5,14,7,8,11,6,15,9,13)) +# b=2 sensitivity analysis techniques A and B that rate the influence of +# k=20 variables/ input parameters (example taken from Iman and Conover, 1987): +ranking <- rbind(A = 1:20, + B = c(1,3,2,4,16,10,19,12,18,17, + 20,5,14,7,8,11,6,15,9,13)) tdcc(ranking, pearson = TRUE, plot = TRUE) } \references{ -Iman and Conover (1987): A Measure of Top-Down Correlation +R. L. Iman and W. J. Conover, + \emph{A Measure of Top-Down Correlation}, + Technometrics, Vol. 29, No. 3 (Aug., 1987), pp. 351--357. } \author{ Stefan Theers