SSlogis5: self start for five-parameter logistic function

In femiguez/nlraa: Nonlinear Regression for Agricultural Applications

self start for five-parameter logistic function

Description

Self starter for a five-parameter logistic function.

Usage

logis5(x, asym1, asym2, xmid, iscal, theta)

SSlogis5(x, asym1, asym2, xmid, iscal, theta)

Arguments

x	input vector (x)
asym1	asymptotic value for low values of x
asym2	asymptotic value for high values of x
xmid	value of x at which y = (asym1 + asym2)/2 (only when theta = 1)
iscal	steepness of transition from asym1 to asym2 (inverse of the scale)
theta	asymmetry parameter, if it is equal to 1, this is the four parameter logistic

Details

The equation for this function is:

f(x) = asym2 + (asym1 - asym2)/(1 + exp(iscal * (log(x) - log(xmid))))^theta

This is known as the Richards' function or the log-logistic and it is described in Archontoulis and Miguez (2015) - (doi:10.2134/agronj2012.0506).

Value

a numeric vector of the same length as x (time) containing parameter estimates for equation specified

logis5: vector of the same length as x (time) using the 5-parameter logistic

Examples

require(ggplot2)
set.seed(1234)
x <- seq(0, 2000, 100)
y <- logis5(x, 35, 10, 800, 5, 2) + rnorm(length(x), 0, 0.5)
dat <- data.frame(x = x, y = y)
fit <- nls(y ~ SSlogis5(x, asym1, asym2, xmid, iscal, theta), data = dat)
## plot
ggplot(data = dat, aes(x = x, y = y)) + 
  geom_point() + 
  geom_line(aes(y = fitted(fit)))

x <- seq(0, 2000)
y <- logis5(x, 30, 10, 800, 5, 2)
plot(x, y)

femiguez/nlraa documentation built on Jan. 26, 2024, 9:31 p.m.