nordpred.estimate: Fit power5 and poisson Age-Period-Cohort models for...

Description Usage Arguments Details Value Note for S-plus Author(s) References See Also Examples

Description

nordpred.estimate estimates parameters in the power5 or poisson Age-Period-Cohort (APC) model

Usage

1
2
3
4
nordpred.estimate(cases, pyr, noperiod, startestage, linkfunc = "power5",
  find_best_model = FALSE)

is.nordpred.estimate(x)

Arguments

cases

A data.frame with number of cases

pyr

A data.frame with observed and forecasted person years.

noperiod

The number of periods to be used in prediction base.

startestage

Youngest age group to include in the regression model

linkfunc

Link function to use in the model. Default is special version used in the Nordpred project ('power5'), where the link is g(x) = x^0.2, while the alternative is the poisson function ('poisson'), where the link is g(x) = log(x)

find_best_model

Logical

x

an object to test for class nordpred

Details

For details of the choice of prediction base, significance test for using recent slope, and for the power5 model, see Moller B., Fekjaer H. et al. (2002), see references.

Value

nordpred.estimate returns an object of class nordpred.estimate (see nordpred.estimate.object).

is.nordpred.estimate returns TRUE if input object is of class nordpred.estimate, FALSE otherwise.

Note for S-plus

Powerlink is made via a special modification in S-PLUS. This works fine for the point estimates, but the variance estimates found via the glm-objects are wrong. For variance estimates, we would rather recommend using R.

Author(s)

Harald Fekjaer and Bjorn Moller (Cancer Registry of Norway)

References

See Also

Other nordpred: is.nordpred, nordpred; nordpred.estimate.object; nordpred.getpred; nordpred.object; nordpred.prediction; plot.nordpred; print.nordpred.estimate; print.nordpred; summary.nordpred

Examples

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
# data (Colon cancer for Norwegian males)
indata

# Create dataset with observed and predicted population
inpop <- cbind(inpop1, inpop2)

# Fit model using powerlink (default):
est <- nordpred.estimate(cases = indata, pyr = inpop, noperiod = 4, startestage = 5)

# Fit model using poisson link:
est2 <- nordpred.estimate(indata, inpop, 4, 5, linkfunc = 'poisson')

# Print results:
print(est)
print(est$glm)

#' # Use estimat object to make predictions:
res <- nordpred.prediction(est, startuseage = 6, cuttrend = c(0, .25, .5, .75, .75), recent = TRUE)

haraldwf/nordpred documentation built on May 17, 2019, 2:41 p.m.