Skip to contents

Obtain Coefficients from Scaled SVD of HMD Data

coef_hmd.Rd

Obtain time series of coefficients from a scaled SVD of data from the Human Mortality Database. The coefficients are a scaled version of the \(U\) matrix from the SVD.

Usage

coef_hmd(
  zipfile,
  n_comp = 5,
  date = c("2025-09-25", "2024-02-26"),
  year_min = NULL,
  eps = 1e-05
)

Arguments

zipfile

The name of a zipped file downloaded from the Human Mortality Database. A path name that is handled by utils::unzip().

n_comp

Number of SVD components to include in result. Default is 5.

date

Date on which data produced. Choices are "2025-09-25", "2024-02-26". (Class is character, not Date.)

year_min

Only include data for year_min onwards. Ignored if year_min is NULL (the default).

eps

Floor for rates. Default is 0.00001.

Value

A tibble

Truncation and transformation

Rates below eps are shifted to eps, and all rates are log-transformed, before the SVD is applied.

See also

  • data_ssvd_hmd() Put data from the Human Mortality Database into the format required by the ssvd() function in bage

  • tidy_hmd() Format data from the Human Mortality Database into a tidy data frame

Examples

zipfile <- system.file("extdata", "hmd_statistics_20250925_subset.zip",
                       package = "bssvd")
coef_hmd(zipfile)
#> # A tibble: 1,515 × 5
#>    sex   country  time component    coef
#>    <fct> <chr>   <int> <chr>       <dbl>
#>  1 Total AUS      1921 Component 1 -1.66
#>  2 Total AUS      1922 Component 1 -1.52
#>  3 Total AUS      1923 Component 1 -1.59
#>  4 Total AUS      1924 Component 1 -1.51
#>  5 Total AUS      1925 Component 1 -1.45
#>  6 Total AUS      1926 Component 1 -1.49
#>  7 Total AUS      1927 Component 1 -1.47
#>  8 Total AUS      1928 Component 1 -1.49
#>  9 Total AUS      1929 Component 1 -1.48
#> 10 Total AUS      1930 Component 1 -1.27
#> # ℹ 1,505 more rows