Skip to contents

Changelog

Source: NEWS.md

demor 1.0.12

  1. LT(): correction of Lx calculation (there was an error when ax was not equal to half the interval).
  2. asdt(): correction of L_not_i calculation. l_not_i starts from l0 so that it shares the radix with lx.
  3. leslie(), ccm() and plot_pyr(): unequal age intervals (e.g. c(0, 1, 5, 10, ...)) are now unified correctly.
  4. leecart():
    • ktadj = "poisson" corrected (there was a problem with offset selection through stats).
    • ktadj = "edaggermin" minimized the squared deviance over a wide interval, and since e-dagger is a U-shaped function of kt it could converge on the wrong branch. It now solves for the root nearest to the unadjusted kt and warns when no root exists.
    • zero mx values are now interpolated log-linearly over age within each year.
    • bx_method = "rotate" now follows Li et al. (2013): the rotation weight uses the constant denominator of their eq. (2), and K(t) is re-fitted as in their eq. (4). The rotation therefore changes the projected age pattern of mortality decline, not the projected e0, so ex0 is now identical for "classic" and "rotate" and the difference shows up only in mx.
  5. decomp(): method = "pollard" is corrected.
  6. yll(): a user-supplied standard no longer produces NA with age.int = 5.
  7. edagger(): correction for the value for the open-ended age group.

demor 1.0.11

  1. fert.approx():
    • now returns an object of class fert.approx
    • added predict.fert.approx(), so fitted fertility schedules can be predicted on new ages with predict(fit, age = 15:49)
    • Added asymptotic uncertainty estimates for fert.approx() via se.method = "asymptotic", which is now the default when se = TRUE. Bootstrap uncertainty estimates remain available with se.method = "bootstrap".
  2. yll() and mort.approx() now use match.arg() for the type and model arguments, so defaults work when the argument is omitted.
  3. edagger(): corrected the contribution of the open (last) age interval, which previously made e-dagger scale with the life-table radix l0 (it biased overall estimate unfortunately). Results are now radix-invariant.
  4. leslie(): correction to the documentation: before defaults were stated incorrectly.

demor 1.0.10

CRAN release: 2026-05-22

  • First CRAN release.