2.8-0 (11/15/2024) ----- When searching for k to match mx to given e0, the ranges are modified if the initial ranges are not wide enough. Default arguments of mortcast.blend change depending if nx = 1 or 5. The pmd functions automatically truncate mx if the age goes beyond ages for which coefficients are available, i.e. older than 100. 2.7-0 (03/31/2022) ----- Added function modpmd that allows to estimate the initial log(mx) in PMD as a_x in Lee-Carter, i.e. it allows to define indices of time periods to be used in the estimation, as well as it allows smoothing. This is propagated into the function copmd, which now accepts arguments use.modpmd, ax.index, ax.smooth and ax.smooth.df. Added argument ax.smooth.df to leecarter.estimate() for passing the degree of freedom for ax smoothing. 2.6-1 (10/15/2021) ----- The updated MLTs converted from tibble to data.frame as it caused problems in reverse dependencies. 2.6-0 (10/08/2021) ----- Modified the CD and UN life tables in datasets MLTlookup and MLT1Ylookup [provided by Sara Hertog, UN]. 2.5-0 (08/06/2021) ----- Function mortcast.blend has an argument match.e0. If TRUE (default) the mx is scaled so that the target e0 matches the input e0. 2.4-0 (06/23/2021) ----- Added a0rule argument to life table functions for computing a0. The default is now the Andreev-Kingkade method (instead of Coale-Demeany). [thanks to Tim Riffe and his DemoTools] Computation of sx for the open age group in life tables for 1-year age groups modified [thanks to Tim Riffe]. Added examples of mx forecasting for one-year age groups. 2.3-0 (05/28/2020) ----- The MLT and PMD methods work now with mx defined for 1-year age groups. 2.2-0/1 (05/01/2020) ----- mortcast() works now with mx defined for 1-year age groups. 2.1-2 (09/27/2019) ----- Bug fix in computing life table's ax for old ages. 2.1-1 (06/10/2019) ----- The abridged life table's ax values for ages 5 and 10 changed to 2.5 due to compatibility with the UN computations. Adaptive sex-ratio adjustments for copmd() (argument adjust.sr.if.needed). 2.1-0 (04/11/2019) ----- New method for projecting mortality: - logquad and logquadj (implements Log-Quadratic Model of Wilmoth et al. 2012) Bug fix in mltj (thanks to Danan Gu) Improvements in the documentation (thanks to Danan Gu and Anupreet Porwal) In the life.table() function, sex option changed from "both" to "total". 2.0-1 (03/19/2019) ----- Added methods for projecting mortality (jointly with Patrick Gerland and Danan Gu): - pmd() and its coherent version copmd() - mlt() and mltj() - mortcast.blend() for combining two methods Support for computing life tables, including one-year age groups life tables: - life.table() 1.2-1 (09/19/2018) ----- Added a constraint to lx in life tables to be positive (thanks to Julia A. Palacios) Bug fix: Argument keep.lt = TRUE in the mortcast function now works. 1.2-0 (05/25/2018) ----- Bug fix (thanks to Julia A. Palacios and Ivan M. Guevara): - The denominator in averaging ax was T+1. Now corrected to T. As a result, k(t) now sums to zero if ax is averaged over all time periods. 1.1-1 (12/18/2017) ----- Added reference. 1.1-0 (12/08/2017) ----- Improvements thanks to suggestions of Marius Pascariu: Functions kannisto.estimate and cokannisto.estimate now also return fitted values and residuals. Fixed construction of the resulting matrix in functions kannisto and cokannisto. New argument in rotate.leecarter, called e0l, giving the starting e0 level of the rotation 1.0-1 (11/30/2017) ----- Initital version.