Fits a coherent functional model to demographic data as described in Hyndman,
Booth & Yasmeen (2012). If two of the series in data are named
male and female, then it will use these two groups. Otherwise
it will use all available groups.
Arguments
- data
demogdata object containing at least two groups.
- order1
Number of basis functions to fit to the model for the geometric mean.
- order2
Number of basis functions to fit to the models for each ratio.
- ...
Extra arguments passed to
fdm.
Value
A list (of class fdmpr) consisting of two objects:
product (an fdm object containing a del for the
geometric mean of the data) and ratio (a list of fdm
objects, being the models for the ratio of each series with the geometric mean).
References
Hyndman, R.J., Booth, H., and Yasmeen, F. (2012) Coherent mortality forecasting: the product-ratio method with functional time series models. Demography, to appear. https://robjhyndman.com/publications/coherentfdm/
Examples
fr.short <- extract.years(fr.sm,1950:2006)
fr.fit <- coherentfdm(fr.short)
summary(fr.fit)
#> *** PRODUCT MODEL ***
#> Functional demographic model
#>
#> Call: fdm(data = rate.product, series = "product", order = order1)
#>
#> Region: FRATNP
#> Data type: mortality
#> Years in fit: 1950 - 2006
#> Ages in fit: 0 - 110
#>
#> Order: 6
#> Percentage variation due to basis functions: 95.1% 2.4% 1.3% 0.4% 0.3% 0.2%
#>
#> Averages across ages:
#> ME MSE MPE MAPE
#> 0.00012 0.00043 -0.00341 0.02145
#>
#> Averages across years:
#> IE ISE IPE IAPE
#> 0.01507 0.04514 -0.16071 2.07038
#>
#>
#> *** RATIO MODELS ***
#>
#> FEMALE
#> Functional demographic model
#>
#> Call: fdm(data = rate.ratio[[j]], series = names(data$rate)[j], order = order2)
#>
#> Region: FRATNP
#> Data type: mortality
#> Years in fit: 1950 - 2006
#> Ages in fit: 0 - 110
#>
#> Order: 6
#> Percentage variation due to basis functions: 79.2% 6.9% 4.9% 2.6% 1.7% 1.1%
#>
#> Averages across ages:
#> ME MSE MPE MAPE
#> 0.00000 0.00015 0.00182 0.07358
#>
#> Averages across years:
#> IE ISE IPE IAPE
#> -0.00041 0.01629 0.25759 8.19290
#>
#>
#> MALE
#> Functional demographic model
#>
#> Call: fdm(data = rate.ratio[[j]], series = names(data$rate)[j], order = order2)
#>
#> Region: FRATNP
#> Data type: mortality
#> Years in fit: 1950 - 2006
#> Ages in fit: 0 - 110
#>
#> Order: 6
#> Percentage variation due to basis functions: 79.2% 6.9% 4.9% 2.6% 1.7% 1.1%
#>
#> Averages across ages:
#> ME MSE MPE MAPE
#> 0.00000 0.00015 0.00182 0.07358
#>
#> Averages across years:
#> IE ISE IPE IAPE
#> 0.00041 0.01629 0.25759 8.19290
#>
plot(fr.fit$product, components=3)
