inst/ubinc/templates/r_analysis_estimation.R

#clearing the workspace
rm(list=ls())
graphics.off()
options(show.error.locations = TRUE)

# If we cannot load the ubiquity package we try the stand alone distribution
if("ubiquity" %in% rownames(installed.packages())){require(ubiquity)} else 
{source(file.path("library", "r_general", "ubiquity.R")) }

# -------------------------------------------------------------------------

# flowctl       = "plot previous estimate"
# flowctl       = "previous estimate as guess"
# flowctl       = "estimate"
  flowctl       = "plot guess"
analysis_name   = "ANAME"
archive_results = TRUE 

# For documentation explaining how to modify the commands below see
# the estimation vignette:
# vignette(package="ubiquity", topic="Estimation")
# Or the estimation tutorial at the bottom of this page:
# http://r.ubiquity.tools/

# -------------------------------------------------------------------------
# Rebuilding the system (R scripts and compiling C code)
cfg = build_system(system_file="<SYSTEM_FILE>",
                   output_directory     = file.path(".", "output"),
                   temporary_directory  = file.path(".", "transient"))


# Initializing the log file ./transient/ubiquity.log
cfg = system_log_init(cfg)
# -------------------------------------------------------------------------

<PSETS>

# # This will estiamte all parameters:
# cfg = system_select_set(cfg, "default")

# # To fix parameters, simply specify only those you want to estimate here:
# pnames = c("PNAME1", "PNAME2")
# cfg = system_select_set(cfg, "default", pnames)
 
# Change initial guess for parameter PNAME to VALUE. The lower bound (lb) and
# upper bound (ub) can also be changed as well
#
# cfg = system_set_guess(cfg, pname="PNAME", value=VALUE, lb=NULL, ub=NULL) 

# -------------------------------------------------------------------------
# Setting options
# 
# Specify output times here using sparse sampling (large time steps) to make
# the estimation quick. See down below where you can specify the sampling to 
# generate smooth profiles when plotting. This will be the
# default output times unless overwritten at the cohort level:
# cfg=system_set_option(cfg, group  = "simulation", 
#                            option = "output_times", 
#                            seq(0,100,1))
#
# cfg=system_set_option(cfg,group = "simulation", option = "solver", value = "lsoda")
#
# To overwrite solver options use the following:
# cfg=system_set_option(cfg,group  = "solver",
#                           option = "atol",   
#                           value  = 1e-10)
#
# cfg=system_set_option(cfg,group  = "solver",
#                           option = "rtol",   
#                           value  = 1e-10)
#
# Uncomment to specify ode file to use. r-file or c-file
# cfg=system_set_option(cfg, group  = "simulation", 
#                            option = "integrate_with",
#                            value  = "r-file")
# 
# Optimizer options
# 
# See the documentation for optim (?optim) for valid values for method and
# elements for the control. You can change the optimizer to optimx and then
# you would just need to adjust the method and control options accordingly.
# The commented code below represents the default values that are used. To
# perform a simulated annealing change the "method" to SANN. 
#
# cfg = system_set_option(cfg, group  = "estimation",
#                              option = "optimizer", 
#                              value  = "optim")
#
# cfg = system_set_option(cfg, group  = "estimation",
#                              option = "method",
#                              value  = "Nelder-Mead")
# 
# cfg = system_set_option(cfg, group  = "estimation",
#                              option = "control", 
#                              value  = list(trace  = TRUE,
#                                            maxit  = 500,
#                                            REPORT = 10))
#
# To use the global optimization routines, uncomment the appropriate lines
# below. IMPORTANT: Because of how the parameter space is sampled it is
# important to set the parameter bounds to reasonable values>
#
# To use particle swarm optimization use the following:
#
# cfg = system_set_option(cfg, group  = "estimation",
#                              option = "optimizer", 
#                              value  = "pso")
# 
# cfg = system_set_option(cfg, group  = "estimation",
#                              option = "method",
#                              value  = "psoptim")
#
# The following will use the genetic algorithm:
#
# cfg = system_set_option(cfg, group  = "estimation",
#                              option = "optimizer", 
#                              value  = "ga")
# 
# cfg = system_set_option(cfg, group  = "estimation",
#                              option = "method",
#                              value  = "ga")
# 
# cfg = system_set_option(cfg, group  = "estimation",
#                              option = "control", 
#                              value  = list(maxiter   = 10000, 
#                                            optimArgs = list(method  = "Nelder-Mead",
#                                                             maxiter = 1000)))


# -------------------------------------------------------------------------
# Loading Datasets
#
# From Excel sheet
# cfg = system_load_data(cfg, dsname     = "DSNAME", 
#                             data_file  = "DS.xls", 
#                             data_sheet = "SHEET")
#
# From csv 
# cfg = system_load_data(cfg, dsname     = "DSNAME", 
#                             data_file  = "DS.csv")
#
# From tab 
# cfg = system_load_data(cfg, dsname     = "DSNAME", 
#                             data_file  = "DS.tab")
#    


# -------------------------------------------------------------------------
# Defining the cohorts
#
# Clearing all of the cohorts
cfg = system_clear_cohorts(cfg)
 
# One entry for each cohort:
# For more information type:
#
# help system_define_cohort
#
# It is necessary to replace the following components: 
#
# CHNAME    - cohort name
# COLNAME   - column name in dataset
# ONAME     - output name, you can use a more descriptive name for your output
# TIMECOL   - column name in dataset with the observation times
# TS        - model timescale corresponding to TIMECOL
# OBSCOL    - column name in dataset with the observation values
# MODOUTPUT - model output corresponding to OBSCOL
#
# Only specify bolus and infusion inputs that are non-zero. Simply ignore
# those that don't exist for the given cohort. Covariates should be specified
#
# The cf elements are optional and will be used to filter the dataset down to
# the records needed for this cohort. If you do not need the placeholders
# simply delete the lines
#
cohort = list(
  name         = "CHNAME",
  cf           = list(COLNAME   = c(),
                      COLNAME   = c()),
  inputs       = NULL,
  outputs      = NULL,
  dataset      = "DSNAME")

<BOLUS><INFUSION_RATES><COVARIATES>


# Here you define information about the outputs you're modeling for this
# cohort. The output "ONAME" used in the list below is a placeholder. You can
# substitute it with a more descriptive string string suchas:
#       PK, Cp_ngml, TV_mm3, etc. 
# No spaces or weird characters.
cohort[["outputs"]][["ONAME"]] = list()
# "of" is the output filter used to further filter the records down to those
# that apply to the current output ONAME. If you do not need them you can just
# delete this part:
cohort[["outputs"]][["ONAME"]][["of"]] = list(
       COLNAME          = c(),
       COLNAME          = c())

# This identifies the columns in the data set that relate to the observation
# times, values and the numerical flag for missing data:
cohort[["outputs"]][["ONAME"]][["obs"]] = list(
         time           = "TIMECOL",
         value          = "OBSCOL",
         missing        = -1)
   
# Here you are mapping model timescales and outputs to the observations. The
# variance model is also specified here. 
#         variance = '1'               # Least squares
#         variance = 'PRED^2'          # Weighted least squares proportional 
#                                      # to the prediction squared
#         variance = '(SLOPE*PRED)^2'  # Maximum likelihood estimation where 
#                                      # SLOPE is defined as a variance parameter <VP>
#        
cohort[["outputs"]][["ONAME"]][["model"]] = list(
         time           = "TS",       
         value          = "MODOUTPUT",
         variance       = "PRED^2")

# This controls the plotted output associated with this cohort
# the color, shape and line values are the values used by ggplot 
# functions. 
cohort[["outputs"]][["ONAME"]][["options"]] = list(
         marker_color   = "black",
         marker_shape   = 16,
         marker_line    = 1 )
cfg = system_define_cohort(cfg, cohort)
# -------------------------------------------------------------------------

# -------------------------------------------------------------------------
# Defining the cohorts with dosing from a NONMEM file
# 
#  DS - name of dataset defined using system_load_dataset
# 
# Include all records in the dataset
#  filter = NULL
# 
# Include only records matching the following filter
#  filter = list()
#  filter$COLNAME = c()
# 
# Mapping information: 
# 
# Inputs:
#  INPUTMAP = list()
#  INPUTMAP$bolus$SPECIES$CMT_NUM            =  1
#  INPUTMAP$infusion_rates$RATE$CMT_NUM      =  1
#  INPUTMAP$covariates$CVNAME$col_COV        = "CNAME"
#
# Outputs
#  OBSMAP = list()
#  OBSMAP$ONAME=list(variance     = "PRED^2",
#                    CMT          =  1,
#                    output       = "<O>",
#                    missing      =  NULL )
#
#  cfg = system_define_cohorts_nm(cfg        =  cfg, 
#                                  filter    =  filter,
#                                  INPUTS    =  INPUTMAP,
#                                  OBS       =  OBSMAP,
#                                  col_GROUP =  NULL,  
#                                  group     =  FALSE)
#
# To overwrite column name mapping the following can be used
#  col_ID    - unique subject identifier 
#  col_CMT   - compartment
#  col_DV    - observations
#  col_TIME  - system time
#  col_AMT   - infusion/dose amounts
#               - these need to be in the same units specified in the system.txt file
#  col_RATE  - rate of infusion or . for bolus
#  col_EVID  - evid (0 - observation, 1 dose)
#  col_GROUP - used to group subjects for plotting purposes. For example if
#              the cohorts were grouped by dose and the column DOSE contained
#              this information, then this will be set to "DOSE".
# 
#  cfg = system_define_cohorts_nm( cfg       = cfg, 
#                                  DS        = "DSNAME",
#                                  col_ID    = "ID",
#                                  col_CMT   = "CMT",
#                                  col_DV    = "DV",
#                                  col_TIME  = "TIME",
#                                  col_AMT   = "AMT",
#                                  col_RATE  = "RATE",
#                                  col_EVID  = "EVID",
#                                  col_GROUP =  NULL,  
#                                  filter    =  filter,
#                                  INPUTS    =  INPUTMAP,
#                                  OBS       =  OBSMAP,
#                                  group     =  FALSE)
#
# -------------------------------------------------------------------------
# performing estimation or loading guess/previous results
pest = system_estimate_parameters(cfg, 
                                  flowctl         = flowctl, 
                                  analysis_name   = analysis_name, 
                                  archive_results = archive_results)

# -------------------------------------------------------------------------

# Here you can specify the output times used for the simulations that will
# be used for plotting. Here you want frequent sampling (small time steps)
# so that you have smoother profiles.
# cfg=system_set_option(cfg, group  = "simulation", 
#                            option = "output_times", 
#                            seq(0,100,1))

# Any cohorts added here will be plotted by will not have been used to
# estimate parameters. This allows you to see how the current estimates
# predict data (validation)

# Simulating the system at the estimates
erp = system_simulate_estimation_results(pest = pest, cfg = cfg) 
# -------------------------------------------------------------------------


plot_opts = c()
# To customize the figures
# Replace ONAME with the specific outputs defined in the cohorts above. Then
# change the values
# plot_opts$outputs$ONAME$yscale   = "linear" # "linear" or "log"
# plot_opts$outputs$ONAME$ylim     = c(0,1)   # NULL
# plot_opts$outputs$ONAME$xlim     = c(0,1)   # NULL
# plot_opts$outputs$ONAME$ylabel   = ""       # output ONAME
# plot_opts$outputs$ONAME$xlabel   = ""       # cohort time units
# plot_opts$tc$width               = 10       # width  of timecourse plot
# plot_opts$tc$height              = 5.5      # height of timecourse plot
# plot_opts$op$width               = 10       # width  of observed vs predicted
# plot_opts$op$height              = 8.0      # height of observed vs predicted

# Plotting the simulated results at the estimates 
# These figures will be placed in output/
system_plot_cohorts(erp, plot_opts, cfg, analysis_name=analysis_name)

#-------------------------------------------------------
# Writing the results to a PowerPoint report
# cfg = system_rpt_read_template(cfg, template="PowerPoint")
# cfg = system_rpt_estimation(cfg=cfg, analysis_name=analysis_name)
# system_rpt_save_report(cfg=cfg, output_file=file.path("output",paste(analysis_name, "-report.pptx", sep="")))
#-------------------------------------------------------
# Writing the results to a Word report
# cfg = system_rpt_read_template(cfg, template="Word")
# cfg = system_rpt_estimation(cfg=cfg, analysis_name=analysis_name)
# system_rpt_save_report(cfg=cfg, output_file=file.path("output",paste(analysis_name, "-report.docx", sep="")))
#-------------------------------------------------------

Try the ubiquity package in your browser

Any scripts or data that you put into this service are public.

ubiquity documentation built on Sept. 5, 2021, 5:15 p.m.