simhetdecay1: Simulate one phase exponential decay data

In TJMurphy/nlfitr: Fitting models to nonlinear data

Description

A sandbox to simulate and visualize random normal heteroscedastic data for a nonlinear decaying response. Variances enlarge with the value of y predicted by the model using a constant coefficeint of variation (cv). The data generating formula is derived from the general model: "y = y0*e^-kx". This model simulates response systems where the rate at which the response decreases is proportional to the level of remaining response. Failure errors can happen in the plot fitting subfunction even though random data is produced. These are more frequent with higher cv values. Just re-simulate with modified parameter values. The regression formula is: 'y ~ (yhi-ylo)*exp(-1*k*x) + ylo'

Usage

simhetdecay1(x, k, ylo, yhi, cv, reps)

Arguments

x	a vector of non-exponential linear scale values representing time.
k	the rate constant, expressed in reciprocal of the X axis time units. The half-life is 0.6932/k.
ylo	the lowest expected y value, or the value at infinite times, expressed in the same units as Y.
yhi	the highest expected y value, or the starting value, expressed in the same units as Y.
cv	the coefficient of variation for y replicates.
reps	an integer value for number of replicates

Value

ggplot, data

Examples

# Note: exponential or log-transformed x scale values will not work
# do not use x = c(1e-9, 3e-9, ...) or c(-9, -8.523, ...)

time <- c(1, 5, 10, 15, 20, 25) # eg, in mins

set.seed(2345)

decay1dat <- simhetdecay1(time, k = 0.15, ylo = 1, yhi = 100, cv = 0.10, reps = 5)


decay1dat$data

TJMurphy/nlfitr documentation built on March 18, 2021, 12:33 p.m.