version 1.1.1

DESCRIPTION

@@ -1,9 +1,9 @@
 Package: WH
 Type: Package
 Title: Enhanced Implementation of Whittaker-Henderson Smoothing
-Version: 1.1.0
+Version: 1.1.1
 Authors@R: person("Guillaume", "Biessy", email = "guillaume.biessy78@gmail.com",
-  role = c("aut", "cre"), comment = c(ORCID = "0000-0003-3756-7345"))
+  role = c("aut", "cre", "cph"), comment = c(ORCID = "0000-0003-3756-7345"))
 Description: An enhanced implementation of Whittaker-Henderson smoothing for the gradation 
     of one-dimensional and two-dimensional actuarial tables used to quantify Life Insurance risks.
     'WH' is based on the methods described in Biessy (2023) <doi:10.48550/arXiv.2306.06932>.
@@ -16,13 +16,15 @@ URL: https://github.com/GuillaumeBiessy/WH
 BugReports: https://github.com/GuillaumeBiessy/WH/issues
 Depends: R (>= 4.2)
 Imports: stats
-Suggests: knitr, rmarkdown, testthat (>= 3.0.0)
+Suggests: knitr, rmarkdown, spelling, testthat (>= 3.0.0)
 Config/testthat/edition: 3
-RoxygenNote: 7.2.3
+RoxygenNote: 7.3.1
 VignetteBuilder: knitr
+Language: en-US
 NeedsCompilation: no
-Packaged: 2023-09-16 11:47:09 UTC; Guillaume BIESSY
-Author: Guillaume Biessy [aut, cre] (<https://orcid.org/0000-0003-3756-7345>)
+Packaged: 2024-03-30 15:51:06 UTC; Guillaume Biessy
+Author: Guillaume Biessy [aut, cre, cph]
+    (<https://orcid.org/0000-0003-3756-7345>)
 Maintainer: Guillaume Biessy <guillaume.biessy78@gmail.com>
 Repository: CRAN
-Date/Publication: 2023-09-18 08:20:02 UTC
+Date/Publication: 2024-04-10 16:40:02 UTC

MD5

@@ -1,52 +1,53 @@
-2e1a1dc73c63f5072005124f7f6ee7ab *DESCRIPTION
+f0b75b6da8f835c34824cee20c5fcc48 *DESCRIPTION
 cda3ab64b68876145fdd6d6032379c16 *NAMESPACE
-f9088b37832cd871673066795553b92b *NEWS.md
-6f005df720dee1e2788140dea3ac0d24 *R/WH-package.R
-cdbd6da5a263f78eef1b2b6e02912118 *R/main.R
+5067dafecc2c1b8b0763bc4d9253423a *NEWS.md
+b133b1090e5677f757a1eb5277a0e63b *R/WH-package.R
+e1b895ecb9e1771ae092709e08dec585 *R/main.R
 4e3a90501d0387bdb1035ebaf7c5501c *R/tools.R
-90cbe261c9ca614a2a18554cfc07b2b4 *README.md
-7c8bfb9d72b020e0c68f9873f61961de *build/vignette.rds
+3f3ff6d6fa13e7d64926ff4dfb19d99e *README.md
+770f7d8a321e201ab7ada04c67addd5a *build/vignette.rds
 43fa807f5c4aa6c8aa8b1301a594e70d *data/portfolio_LTC.rda
 389c4d1f41c027674a25bf81c71fccf8 *data/portfolio_mort.rda
-13e5b5217d0ca0ae468daab576cdd2db *inst/CITATION
+54f678fb74ad9318fb08d433ad877986 *inst/CITATION
+712304661eecd6942e913ba103cfee62 *inst/WORDLIST
 d847721c40983fb15fdb62c0b9e141c8 *inst/doc/WH.R
-b2438c9365b57f8dc207cffc3b229c80 *inst/doc/WH.Rmd
-efba59e111c312606d010e0bd61abbf6 *inst/doc/WH.html
-2d7c65133d7934fd4a6f97fcdab7cc2e *man/WH-package.Rd
-d442149a14024490020343be20420da0 *man/WH_1d.Rd
-889d09be20652fdc6ffb4cc631ce0a5a *man/WH_1d_fixed_lambda.Rd
-6cac4e003cf20fe118935116c371f8cb *man/WH_1d_outer.Rd
-3ccec8ddf5f0423e803659ca4642ecef *man/WH_1d_perf.Rd
+8f7e3bd1c1dfd1fa0c370e3a8e2c5900 *inst/doc/WH.Rmd
+90dd96cc82b175d884a19ad07c302166 *inst/doc/WH.html
+0cedf3628fb7d43de64d546010ac7b68 *man/WH-package.Rd
+c681a22ef91c0a8166e9a04b07b90636 *man/WH_1d.Rd
+87c11e97582de438f0f643b91ae700d9 *man/WH_1d_fixed_lambda.Rd
+d6bf463db5dbe034adb2fb2b4814cc35 *man/WH_1d_outer.Rd
+4d300d076450ab1273c4e546ba7f7ca5 *man/WH_1d_perf.Rd
 357d0d5c34cc66b4e095c0ca538f79d2 *man/WH_2d.Rd
-4ab1a2780f9ae7126464586d7d6577e7 *man/WH_2d_fixed_lambda.Rd
-59cbf4c1489a31144a94c1bac0297795 *man/WH_2d_outer.Rd
-4ab6d6eadc694865216157d2cd4d539e *man/WH_2d_perf.Rd
+c30efae626f31a93a96c5dfda631cb19 *man/WH_2d_fixed_lambda.Rd
+637eac1756b5a21951538762157e9839 *man/WH_2d_outer.Rd
+9754652bb5cf72e62534e5fba64583be *man/WH_2d_perf.Rd
 e718b63482111fbb469b0d7a6eb447ca *man/blockdiag.Rd
-495e497b6cb6e4dd81b8377fa85a521f *man/build_D_mat.Rd
-3797917646aa550a487c894ef5123562 *man/compute_res_deviance.Rd
-f474d77fcdc426cc29ac5fe5475fbbbf *man/eigen_dec.Rd
+19ba916745c2273fee313d3a6ae3cdcd *man/build_D_mat.Rd
+655323575573de23a5c40bc1ef3294ad *man/compute_res_deviance.Rd
+c389e1b0e35b7efc0f895358c3f4b17a *man/eigen_dec.Rd
 eaef6ac0f3b6e7b13ffdf2e0af009a11 *man/figures/README-plot-1.png
 c9ba1845c022a3033eb0181914837eac *man/figures/README-plot-2.png
 322acdd22207c3b58d5b964ada6026d3 *man/figures/README-plot-3.png
 a97393938e4803c65f958748c3d3b49a *man/figures/README-plot-4.png
 7e48b945d48d76c2c88dd6bedd6138d5 *man/figures/README-plot-5.png
 464d10356c7b350846c2b1a04227ce90 *man/figures/README-predict-1.png
 04fe2b0f7cfaa7d6f7d4d4a27a424c3f *man/figures/README-predict-2.png
-4e6e1418d443718c7de38fce6036de54 *man/get_diagnosis.Rd
-d5e82535ed8b3fceab4cc7509b156a4b *man/map.Rd
-c3ea0d729406c8ed3ddfd2694d8b46ae *man/map2.Rd
-06f6848a69722cf63ab62ac3da08f89b *man/output_to_df.Rd
-8a1b07ebc43dbc50a23939802275b475 *man/plot.WH_1d.Rd
-bd29980fbd8d37f235cf65a15555308a *man/plot.WH_2d.Rd
-283d764e69419d9dcba4233388ceea48 *man/portfolio_LTC.Rd
-c2269a20aa5c5f6839e2a2cf9a355d04 *man/portfolio_mort.Rd
-83455467c459d81b4cc7ed8952fe3a65 *man/predict.WH_1d.Rd
+6f2f320f3946cbb76e0f0099ea4aeee4 *man/get_diagnosis.Rd
+e19a925c5ec6459c97c66ff91fb12a64 *man/map.Rd
+f4d10b90920ef1541549e7f1036dafd8 *man/map2.Rd
+91b8bdf7bfee88584e9fde5ec02a535b *man/output_to_df.Rd
+102c1700bbfad0c56ff69e7debb5aeaa *man/plot.WH_1d.Rd
+03806d2b9477eb8de24e893055a3b70b *man/plot.WH_2d.Rd
+b6b8cc174374e3cbecd75761864cef86 *man/portfolio_LTC.Rd
+7a2d7205f04fae4319cb1e5d75fc822a *man/portfolio_mort.Rd
+b4d07169f108d39f74ad3dc610d07eb4 *man/predict.WH_1d.Rd
 b82976558c28dc55d3cc2a4562b7faf9 *man/predict.WH_2d.Rd
-9a98438678dfc5b9ee8d219f97e5515e *man/print.WH_1d.Rd
-c4ed3089c4af14ec341cdcaa83e2841c *man/print.WH_2d.Rd
+b2865b8db63c0d26a7c26949ea3ddcb7 *man/print.WH_1d.Rd
+b654d5a850d9572005cab559927c5119 *man/print.WH_2d.Rd
+332173f20d20942c02019350ea4dcadc *tests/spelling.R
 828d0a66030e7522cf399e9e2c90bca4 *tests/testthat.R
-98b0b2097af4a4ed1574e915654f2ed5 *tests/testthat/test-1D.R
-e430db3a04daae6211dac97fd26ca510 *tests/testthat/test-2D.R
-b2438c9365b57f8dc207cffc3b229c80 *vignettes/WH.Rmd
-ecfe2ac0176df273cadfa9e69fdb6815 *vignettes/Whittaker_Henderson_plots.R
+dbb34774c5dc559070651d626ded7e1a *tests/testthat/test-1D.R
+4b8cf2f662743c39bc58b3b108854fa9 *tests/testthat/test-2D.R
+8f7e3bd1c1dfd1fa0c370e3a8e2c5900 *vignettes/WH.Rmd
 5f5187a8555f27efd51e5cfc43d0a846 *vignettes/biblio.bib

NEWS.md

@@ -1,45 +1,49 @@
-# WH 1.1.0
-
-* Made significant changes to the extrapolation methods provided by be `predict.WH_1d` and `predict.WH_2d` functions. It turns out that the formula used for the extrapolation ignored the innovation error caused by the prior on the extrapolated region, resulting in smaller credibility intervals than they should have been. This now has been fixed.
-
-# WH 1.0.7
-
-* Removed tests of the form `expect_equal(f(x), f(x))` after confirmation they do not work with MKL BLAS
-
-# WH 1.0.6
-
-* Further improved test robustness
-
-* Added tests of the form `expect_equal(f(x), f(x))` for testing purposes
-
-# WH 1.0.5
-
-* Further improved tests robustness
-
-# WH 1.0.4
-
-* Simplified computation of the matrix `tUWU` by using the `crossprod` function, which should reduce memory usage (by half) and computation time (slightly)
-
-* Fixed an issue with the `WH_2d` plot and the what = "edf" argument
-
-* Improved tests robustness
-
-# WH 1.0.3
-
-* Replaced backquotes by normal quotes in the description field of the DESCRIPTION file
-
-# WH 1.0.2
-
-* Updated the package description to link it to the paper whose methods are used in the package
-
-# WH 1.0.1
-
-* Fixed an issue in the author field of the DESCRIPTION file
-
-# WH 1.0.0
-
-* Official candidate for CRAN release
-
-# WH 0.1.0
-
-* Added a `NEWS.md` file to track changes to the package
+# WH 1.1.1
+
+* Increased tolerance to remove a lone failing test which only occurs when using MKL BLAS
+
+# WH 1.1.0
+
+* Made significant changes to the extrapolation methods provided by be `predict.WH_1d` and `predict.WH_2d` functions. It turns out that the formula used for the extrapolation ignored the innovation error caused by the prior on the extrapolated region, resulting in smaller credibility intervals than they should have been. This now has been fixed.
+
+# WH 1.0.7
+
+* Removed tests of the form `expect_equal(f(x), f(x))` after confirmation they do not work with MKL BLAS
+
+# WH 1.0.6
+
+* Further improved test robustness
+
+* Added tests of the form `expect_equal(f(x), f(x))` for testing purposes
+
+# WH 1.0.5
+
+* Further improved tests robustness
+
+# WH 1.0.4
+
+* Simplified computation of the matrix `tUWU` by using the `crossprod` function, which should reduce memory usage (by half) and computation time (slightly)
+
+* Fixed an issue with the `WH_2d` plot and the what = "edf" argument
+
+* Improved tests robustness
+
+# WH 1.0.3
+
+* Replaced backquotes by normal quotes in the description field of the DESCRIPTION file
+
+# WH 1.0.2
+
+* Updated the package description to link it to the paper whose methods are used in the package
+
+# WH 1.0.1
+
+* Fixed an issue in the author field of the DESCRIPTION file
+
+# WH 1.0.0
+
+* Official candidate for CRAN release
+
+# WH 0.1.0
+
+* Added a `NEWS.md` file to track changes to the package

R/WH-package.R

@@ -7,9 +7,8 @@
 #' original smoothing algorithm to maximum likelihood estimation, automatically
 #' selects the smoothing parameter(s) and extrapolates beyond the range of data.
 #'
-#' @docType package
 #' @name WH-package
-NULL
+"_PACKAGE"
 
 #' Agregated Mortality Dataset
 #'

R/main.R

@@ -27,7 +27,7 @@
 #' @param method Method to be used to find the optimal smoothing parameter.
 #'   Default to `"fixed_lambda"` if `lambda` is supplied, meaning no
 #'   optimization is performed. Otherwise, default to the most reliable
-#'   `"outer"` methodbased on the `optimize` function from package `stats`.
+#'   `"outer"` method based on the `optimize` function from package `stats`.
 #' @param q Order of penalization. Polynoms of degrees `q - 1` are considered
 #'   completely smooth and are therefore unpenalized. Should be left to the
 #'   default of `2` for most practical applications.

README.md

@@ -39,7 +39,7 @@ are mandatory for those functions:
   two-dimension case) whose dimensions should match those of `d`. The
   contribution of each individual to the portfolio central exposure
   corresponds to the time the individual was actually observed with
-  corresponding age (and duration in the two-dimension cas). It always
+  corresponding age (and duration in the two-dimension case). It always
   ranges from 0 to 1 and is affected by individuals leaving the
   portfolio, no matter the cause, as well as censoring and truncating
   phenomena.

inst/CITATION

@@ -2,5 +2,6 @@ bibentry(
   bibtype  = "Unpublished",
   title    = "Revisiting Whittaker-Henderson Smoothing",
   author   = person("Guillaume Biessy"),
-  note     = "Submitted to Insurance: Mathematics and Economics in June 2023"
+  note     = "Working Paper",
+  year     = 2023
 )

inst/WORDLIST

@@ -0,0 +1,66 @@
+Agregated
+BLAS
+CMD
+Carballo
+Dae
+Diag
+Eigen
+Fellner
+Jin
+LTC
+Lapply
+MKL
+ORCID
+Polynoms
+Schall
+TP
+TW
+Whittaker
+agregated
+agregating
+arXiv
+argmax
+argmin
+bidimensional
+bmatrix
+boldsymbol
+cas
+ddots
+diag
+doi
+durations
+ec
+edf
+eigen
+exponentiating
+frac
+ggplot
+inputed
+invertible
+lbrace
+ldots
+le
+leq
+ln
+mathbb
+mathbf
+mathcal
+mathsf
+ne
+neq
+odot
+optim
+otimes
+overset
+pmatrix
+posteriori
+propto
+purrr
+rbrace
+textbf
+underset
+unidimensional
+vdots
+vec
+vectorized
+widehat

inst/doc/WH.Rmd

@@ -197,7 +197,7 @@ WH_2d_df <- WH_2d_fit |> output_to_df()
 
 ### How to obtain credibility intervals?
 
-From the explicit solution of WH smoothinh, $\mathbb{E}(\hat{\mathbf{y}}) = (W + P_{\lambda})^{-1}W\mathbb{E}(\mathbf{y}) \ne \mathbb{E}(\mathbf{y})$ when $\lambda \ne 0$. This implies that penalization introduces a smoothing bias, which prevents the construction of a confidence interval centered around $\mathbb{E}(\mathbf{y})$. Therefore, in this section, we turn to a Bayesian approach where smoothing can be interpreted more naturally.
+From the explicit solution of WH smoothing, $\mathbb{E}(\hat{\mathbf{y}}) = (W + P_{\lambda})^{-1}W\mathbb{E}(\mathbf{y}) \ne \mathbb{E}(\mathbf{y})$ when $\lambda \ne 0$. This implies that penalization introduces a smoothing bias, which prevents the construction of a confidence interval centered around $\mathbb{E}(\mathbf{y})$. Therefore, in this section, we turn to a Bayesian approach where smoothing can be interpreted more naturally.
 
 Let us suppose that $\mathbf{y} | \boldsymbol{\theta}\sim \mathcal{N}(\boldsymbol{\theta}, W^{-})$ and $\boldsymbol{\theta} \sim \mathcal{N}(0, P_{\lambda}^{-})$ where $P_{\lambda}^{-}$ denotes the pseudo-inverse of the matrix $P_{\lambda}$. The Bayes' formula allows us to express the posterior likelihood $f(\boldsymbol{\theta} | \mathbf{y})$ associated with these choices in the following form: