predict-methods: Predicted values from an mle2 fit

In bbmle: Tools for general maximum likelihood estimation

Description

Given an mle2 fit and an optional list of new data, return predictions (more generally, summary statistics of the predicted distribution)

Usage

    ## S4 method for signature 'mle2'
predict(object, newdata=NULL,
                         location="mean", newparams=NULL, ...)
    ## S4 method for signature 'mle2'
simulate(object, nsim,
                         seed, newdata=NULL, newparams=NULL, ...)
    ## S4 method for signature 'mle2'
residuals(object,type=c("pearson","response"),
                   location="mean",...)

Arguments

object	an mle2 object
newdata	optional list of new data
newparams	optional vector of new parameters
location	name of the summary statistic to return
nsim	number of simulations
seed	random number seed
type	residuals type
...	additional arguments (for generic compatibility)

Methods

x = "mle2"

an mle2 fit

Note

For some models (e.g. constant models), predict may return a single value rather than a vector of the appropriate length.

Examples

set.seed(1002)
lymax <- c(0,2)
lhalf <- 0
x <- runif(200)
g <- factor(rep(c("a","b"),each=100))
y <- rnbinom(200,mu=exp(lymax[g])/(1+x/exp(lhalf)),size=2)
dat <- data.frame(y,g,x)

fit3 <- mle2(y~dnbinom(mu=exp(lymax)/(1+x/exp(lhalf)),size=exp(logk)),
    parameters=list(lymax~g),
    start=list(lymax=0,lhalf=0,logk=0),
data=dat)

plot(y~x,col=g)
## true curves
curve(exp(0)/(1+x/exp(0)),add=TRUE)
curve(exp(2)/(1+x/exp(0)),col=2,add=TRUE)
## model predictions
xvec = seq(0,1,length=100)
lines(xvec,predict(fit3,newdata=list(g=factor(rep("a",100),levels=c("a","b")),
                                x = xvec)),col=1,lty=2)
lines(xvec,predict(fit3,newdata=list(g=factor(rep("b",100),levels=c("a","b")),
                                x = xvec)),col=2,lty=2)


## comparing automatic and manual predictions
p1 = predict(fit3)
p2A =
with(as.list(coef(fit3)),exp(`lymax.(Intercept)`)/(1+x[1:100]/exp(lhalf)))
p2B =
with(as.list(coef(fit3)),exp(`lymax.(Intercept)`+lymax.gb)/(1+x[101:200]/exp(lhalf)))
all(p1==c(p2A,p2B))
##
simulate(fit3)

Example output

Loading required package: stats4
[1] TRUE
  [1]  0  1  0  2  1  0  0  1  0  0  0  1  1  1  2  0  0  0  0  1  1  2  0  0  0
 [26]  1  0  2  0  0  0  1  2  0  1  0  1  0  2  0  1  1  2  2  0  0  0  1  0  0
 [51]  0  2  1  0  0  0  1  0  0  2  1  0  1  0  0  0  0  0  0  0  0  2  2  1  0
 [76]  1  0  1  0  1  1  0  0  0  0  0  0  0  0  0  0  0  1  1  1  1  0  0  0  0
[101]  2 17  0  8  7  1  8  8  2 17  2  2  4  6  1  8  0  9 12  1  2  4 20  0  1
[126]  2  5  2  3  5  4  0 12  8  1  2  0  2  1  3  4  9  2 15 10 13  1  2  6  7
[151]  2  2  4 14  9  2 12  2  4  5 14  1  4  2 17  7  4  8  5  4  2  9  2  4  2
[176]  9  4  1 10  6  1  7  9  7  6  5  2  3  0  3  4  8  1  3 24  1  0  0 14  7

bbmle documentation built on May 2, 2019, 4:54 p.m.