From ea6152fa5df0ca9e794d3ed1d196802ff7d48667 Mon Sep 17 00:00:00 2001
From: Ioannis Kosmidis <ioannis.kosmidis@warwick.ac.uk>
Date: Wed, 8 Feb 2023 12:53:07 +0000
Subject: [PATCH 1/7] Bumb to development version

---
 DESCRIPTION | 2 +-
 NEWS.md     | 2 ++
 2 files changed, 3 insertions(+), 1 deletion(-)

diff --git a/DESCRIPTION b/DESCRIPTION
index 4035eb1..2522b27 100644
--- a/DESCRIPTION
+++ b/DESCRIPTION
@@ -1,6 +1,6 @@
 Package: brglm2
 Title: Bias Reduction in Generalized Linear Models
-Version: 0.9.1
+Version: 0.9.999
 Authors@R: c(person(given = "Ioannis", family = "Kosmidis", role = c("aut", "cre"), email = "ioannis.kosmidis@warwick.ac.uk", comment = c(ORCID = "0000-0003-1556-0302")),
 	     person(given = c("Euloge", "Clovis"), family = c("Kenne Pagui"), role = "aut", email = "kenne@stat.unipd.it"),
 	     person(given = "Kjell", family = "Konis", role = "ctb", email = "kjell.konis@me.com"),
diff --git a/NEWS.md b/NEWS.md
index 3bb871b..4ee691a 100644
--- a/NEWS.md
+++ b/NEWS.md
@@ -1,3 +1,5 @@
+# brglm2 0.9.999
+
 # brglm2 0.9.1
 
 ## Other improvements, updates and additions

From f19a1c9a370418d43f8ef7723ddce80a62ee2949 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Staffan=20Betn=C3=A9r?= <staffan@betner.nu>
Date: Tue, 29 Aug 2023 15:53:25 +0200
Subject: [PATCH 2/7] Small typo corrected

---
 vignettes/iteration.Rmd | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/vignettes/iteration.Rmd b/vignettes/iteration.Rmd
index 143ff2a..c243373 100755
--- a/vignettes/iteration.Rmd
+++ b/vignettes/iteration.Rmd
@@ -144,7 +144,7 @@ The results in @kosmidis:2019 can be used to show that if
 \end{align*}
 then the solution of the adjusted score equations $s_\beta + A_\beta = 0_p$ and $s_\phi + A_\phi = 0$ results in estimators $\tilde\beta$ with mean bias of small asymptotic order than the maximum likelihood estimator and $\tilde\phi$ with median bias of smaller asymptotic order than the maximum likelihood estimator.
 
-### Maximum penalized likelihood with powers of Jefreys prior as penalty
+### Maximum penalized likelihood with powers of Jeffreys prior as penalty
 The likelihood penalized by a power of the Jeffreys prior
 \[
 |i_{\beta\beta}|^a |i_{\phi\phi}|^a \quad a > 0

From bdc4ced5b20274c0635e8f324d8fdf4bd0394961 Mon Sep 17 00:00:00 2001
From: Ioannis Kosmidis <ioannis.kosmidis@warwick.ac.uk>
Date: Thu, 5 Oct 2023 11:49:36 +0100
Subject: [PATCH 3/7] Updates on convergence criteria + code base + docs +
 argument checks

---
 NEWS.md                             | 54 ++++++++++++++++-------------
 R/brglm2-package.R                  |  1 +
 R/brglmControl.R                    |  5 +++
 R/brglmFit.R                        | 41 ++++++++++++----------
 inst/tinytest/test-binomial.R       |  2 +-
 inst/tinytest/test-bracl.R          |  8 ++---
 inst/tinytest/test-jeffreys.R       |  2 +-
 inst/tinytest/test-transformation.R |  2 +-
 man/brglm2.Rd                       |  1 +
 9 files changed, 66 insertions(+), 50 deletions(-)

diff --git a/NEWS.md b/NEWS.md
index 4ee691a..31e9e89 100644
--- a/NEWS.md
+++ b/NEWS.md
@@ -1,8 +1,16 @@
-# brglm2 0.9.999
+# brglm2 0.9.2
+
+## Improvements, updates and additions
+
+* Convergence of the `brglm_fit` iterations is now determined if the L^Inf norm of the step size (rather than the L^1 as it was previously) of the quasi-Fisher scoring procedure is less than `epsilon` (see `?brglm_control` for the definition of `epsilon`). This is more natural as `epsilon` then determines directly the precision of the reported estimates and does not depend on their number.
+
+* `brglm_control()` now checks that the supplied value of `max_step_factor` is numeric and greater or equal to `1`. If not, then it is set to the default value of `12`.
+
+
 
 # brglm2 0.9.1
 
-## Other improvements, updates and additions
+## Improvements, updates and additions
 
 * Added the `enzymes` and `hepatitis` data sets (from the [*pmlr*](https://cran.r-project.org/package=pmlr)) to support examples and tests.
 
@@ -16,7 +24,7 @@
 
 * Fixed a bug where the dispersion in the resulting object would not be transformed even if `transformation != "identity"` when `type` is `ML` or `AS_median` or `AS_mixed`.
 
-## Other improvements, updates and additions
+## Improvements, updates and additions
 
 * Moved unit tests to [**tinytest**](https://cran.r-project.org/package=tinytest).
 
@@ -28,7 +36,7 @@
 
 # brglm2 0.8.2
 
-## Other improvements, updates and additions
+## Improvements, updates and additions
 
 * Housekeeping.
 * Removed lpSolveAPI from imports.
@@ -39,7 +47,7 @@
 
 * Fixed a bug when predicting from `bracl` objects with non-identifiable parameters.
 
-## Other improvements, updates and additions
+## Improvements, updates and additions
 
 * Work on output consistently from `print()` methods for `summary.XYZ`
   objects; estimator type is now printed and other fixes.
@@ -74,7 +82,7 @@
 * `brglmFit()` iteration returns last estimates that worked if
   iteration fails.
 
-## Other improvements, updates and additions
+## Improvements, updates and additions
 
 * Documentation and example updates.
 
@@ -94,7 +102,7 @@
   (`check_aliasing = FALSE`) rank deficiency checks (through a QR
   decomposition of the model matrix), saving some computational effort.
 
-## Other improvements, updates and additions
+## Improvements, updates and additions
 * updated DOI links in documentation and some http -> https fixes.
 
 # brglm2 0.7.0
@@ -106,7 +114,7 @@
 ## New functionality
 * `confint` method for `brmulitnom` objects
 
-## Other improvements, updates and additions
+## Improvements, updates and additions
 * Updated reference to [Kenne Pagui et al (2017)](https://doi.org/10.1093/biomet/asx046).q
 * Updated reference to [Kosmidis and Firth (2020)](https://doi.org/10.1093/biomet/asaa052).
 * Fixed issues with references.
@@ -114,7 +122,7 @@
 
 # brglm2 0.6.2
 
-## Other improvements, updates and additions
+## Improvements, updates and additions
 * `vcov.brglmFit()` now uses `vcov.summary.glm()` and supports the
   `complete` argument for controlling whether the variance covariance
   matrix should include rows and columns for aliased parameters.
@@ -132,7 +140,7 @@
   `bracl` objects.
 * `detect_separation()` now handles one-column model matrices correctly.
 
-## Other improvements, updates and additions
+## Improvements, updates and additions
 * Documentation improvements and typo fixes.
 
 # brglm2 0.6
@@ -142,7 +150,7 @@
   supported generalized linear models. See the help files of
   `brglmControl()` and `brglmFit()` for details.
 
-## Other improvements, updates and additions
+## Improvements, updates and additions
 * Documentation updates and improvements.
 * Updated vignettes to include maximum penalized likelihood with
   powers of the Jeffreys prior as penalty.
@@ -158,14 +166,14 @@
   for more fine-tuning of the starting values when `brglmFit()` is
   called with `start = NULL`.
 
-## Other improvements, updates and additions
+## Improvements, updates and additions
 * Documentation updates and improvements.
 * Added Kosmidis et al (2019) in the description file.
 * Added tests for `brglmControl()`.
 
 # brglm2 0.5.1
 
-## Other improvements, updates and additions
+## Improvements, updates and additions
 * Fixed typos in vignettes and documentation.
 * Added ORCHID for Ioannis Kosmidis in DESCRIPTION.
 
@@ -193,7 +201,7 @@ Added `residuals()` methods for `brmultinom` and `bracl` objects
   misclassification in binomial response models (Neuhaus, 1999,
   Biometrika).
 
-## Other improvements, updates and additions
+## Improvements, updates and additions
 * Improved `summary()` method for `brmultinom` objects.
 * Better starting values for null fits.
 * Added references to [arxiv:1804.04085](https://arxiv.org/abs/1804.04085) in
@@ -201,7 +209,7 @@ Added `residuals()` methods for `brmultinom` and `bracl` objects
 * Updated reference to [Kenne Pagui et al (2017)](https://doi.org/10.1093/biomet/asx046).
 
 # brglm2 0.1.8
-## Other improvements, updates and additions
+## Improvements, updates and additions
 * Improved documentation examples.
 * Removed warning about observations with non-positive weights from brmultinom.
 * Updated email address for Ioannis Kosmidis in brglmFit.
@@ -214,7 +222,7 @@ Added `residuals()` methods for `brmultinom` and `bracl` objects
 
 # brglm2 0.1.7
 
-## Other improvements, updates and additions
+## Improvements, updates and additions
 * Eliminated errors from markdown chunks in multinomial vignette.
 
 # brglm2 0.1.6
@@ -222,7 +230,7 @@ Added `residuals()` methods for `brmultinom` and `bracl` objects
 ## Bug fixes
 * Compatibility with new version of enrichwith.
 
-## Other improvements, updates and additions
+## Improvements, updates and additions
 * New email for Ioannis Kosmidis.
 
 # brglm2 0.1.5
@@ -239,7 +247,7 @@ Added `residuals()` methods for `brmultinom` and `bracl` objects
   `detect_separation()` methods in line with the update of
   `glm.fit()`.
 
-## Other improvements, updates and additions
+## Improvements, updates and additions
 * less strict tolerance in `brglm_control()`.
 * Updates to help files.
 * Fixed typos in iteration vignette.
@@ -265,7 +273,7 @@ Added `residuals()` methods for `brmultinom` and `bracl` objects
   `detectSeparationControl()`, `detect_separation_control()`,
   `checkInfiniteEstimates()`, `check_infinite_estimates()`).
 
-## Other improvements, updates and additions
+## Improvements, updates and additions
 * Minor enhancements in the codebase.
 * The inverse expected information matrix is computed internally using
   `cho2inv()`.
@@ -274,12 +282,8 @@ Added `residuals()` methods for `brmultinom` and `bracl` objects
 
 # brglm2 0.1.3
 
-## Bug fixes
-
-## New functionality
-
-## Other improvements, updates and additions
-* Fixed typo in f_{Y_i}(y) in iteration vignette (thanks to Eugene
+## Improvements, updates and additions
+* Fixed typo in $f_{Y_i}(y)$ in iteration vignette (thanks to Eugene
   Clovis Kenne Pagui for spotting),
 
 # brglm2 0.1.2
diff --git a/R/brglm2-package.R b/R/brglm2-package.R
index ee23ba0..025a599 100644
--- a/R/brglm2-package.R
+++ b/R/brglm2-package.R
@@ -96,6 +96,7 @@
 #' 185-196. \doi{10.1002/wics.1296}.
 #'
 #' @docType package
+#' @aliases brglm2-package
 #' @name brglm2
 #' @import stats
 #' @import enrichwith
diff --git a/R/brglmControl.R b/R/brglmControl.R
index 8401574..9503412 100644
--- a/R/brglmControl.R
+++ b/R/brglmControl.R
@@ -201,6 +201,11 @@ brglmControl <- function(epsilon = 1e-06, maxit = 100,
     }
     if (!is.numeric(epsilon) || epsilon <= 0)
         stop("value of 'epsilon' must be > 0")
+
+    if (!is.numeric(max_step_factor) || max_step_factor < 1) {
+        warning("`max_step_factor = ", deparse(max_step_factor), "` is not a permissible value. Defaulting to 12")
+        max_step_factor <- 12
+    }
     list(epsilon = epsilon, maxit = maxit, trace = trace,
          check_aliasing = check_aliasing,
          response_adjustment = response_adjustment,
diff --git a/R/brglmFit.R b/R/brglmFit.R
index f36ffeb..3ab1284 100644
--- a/R/brglmFit.R
+++ b/R/brglmFit.R
@@ -876,29 +876,35 @@ brglmFit <- function(x, y, weights = rep(1, nobs), start = NULL, etastart = NULL
             for (iter in seq.int(control$maxit)) {
                 step_factor <- 0
                 testhalf <- TRUE
+
                 ## Inner iteration
                 while (testhalf & step_factor < control$max_step_factor) {
                     ## store previous values
-                    betas0 <- betas; dispersion0 <- dispersion
+                    ## betas0 <- betas
+                    ## dispersion0 <- dispersion
                     step_beta_previous <- step_beta
                     step_zeta_previous <- step_zeta
+
                     ## Update betas
                     betas <- betas + slowit * 2^(-step_factor) * step_beta
+
                     ## Update zetas
                     if (!no_dispersion & df_residual > 0) {
                         transformed_dispersion <- eval(control$Trans)
                         transformed_dispersion <- transformed_dispersion + 2^(-step_factor) * step_zeta
                         dispersion <- eval(control$inverseTrans)
                     }
+
                     ## Compute key quantities
                     theta <- c(betas, dispersion)
                     transformed_dispersion <- eval(control$Trans)
-                    ## Mean quantities
 
+                    ## Mean quantities
                     quantities <- try(key_quantities(theta, y = y, level = 2 * !no_dispersion, scale_totals = has_fixed_totals, qr = TRUE), silent = TRUE)
                     ## This is to capture qr failing and revering to previous estimates
                     if (failed_adjustment_beta <- inherits(quantities, "try-error")) {
-                        betas <- betas0; dispersion <- dispersion0
+                        ## betas <- betas0
+                        ## dispersion <- dispersion0
                         warning("failed to calculate score adjustment")
                         break
                     }
@@ -912,10 +918,9 @@ brglmFit <- function(x, y, weights = rep(1, nobs), start = NULL, etastart = NULL
                         warning("failed to calculate score adjustment")
                         break
                     }
-                    adjusted_grad_beta <- with(step_components_beta, {
-                        grad + adjustment
-                    })
+                    adjusted_grad_beta <- with(step_components_beta, grad + adjustment)
                     step_beta <- drop(step_components_beta$inverse_info %*% adjusted_grad_beta)
+
                     ## Dispersion quantities
                     if (no_dispersion) {
                         adjusted_grad_zeta <- step_zeta <- NA_real_
@@ -929,28 +934,28 @@ brglmFit <- function(x, y, weights = rep(1, nobs), start = NULL, etastart = NULL
                             warning("failed to calculate score adjustment")
                             break
                         }
-                        adjusted_grad_zeta <- with(step_components_zeta, {
-                            grad + adjustment
-                        })
+                        adjusted_grad_zeta <- with(step_components_zeta, grad + adjustment)
                         step_zeta <- as.vector(adjusted_grad_zeta * step_components_zeta$inverse_info)
                     }
+
+                    ## Convergence criteria
+                    linf_current <- max(abs(c(step_beta, step_zeta)), na.rm = TRUE)
+                    linf_previous <- max(abs(c(step_beta_previous, step_zeta_previous)), na.rm = TRUE)
+                    testhalf <- linf_current > linf_previous
+
                     ## Continue inner loop
-                    if (step_factor == 0 & iter == 1)  {
-                        testhalf <- TRUE
-                    } else {
-                        s2 <- c(abs(step_beta), abs(step_zeta))
-                        s1 <- c(abs(step_beta_previous), abs(step_zeta_previous))
-                        testhalf <- sum(s2, na.rm = TRUE) > sum(s1, na.rm = TRUE)
-                    }
+                    ## if (step_factor == 0 & iter == 1)  {
+                    ##     testhalf <- TRUE
+                    ## }
                     step_factor <- step_factor + 1
+
                     ##  Trace here
                     if (control$trace) {
-
                         trace_iteration()
                     }
                 }
                 failed <- failed_adjustment_beta | failed_inversion_beta | failed_adjustment_zeta | failed_inversion_zeta
-                if (failed | sum(abs(c(step_beta, step_zeta)), na.rm = TRUE) < control$epsilon) {
+                if (failed | linf_current < control$epsilon) {
                     break
                 }
             }
diff --git a/inst/tinytest/test-binomial.R b/inst/tinytest/test-binomial.R
index 7197c01..602e683 100644
--- a/inst/tinytest/test-binomial.R
+++ b/inst/tinytest/test-binomial.R
@@ -2,7 +2,7 @@ library("brglm")
 data("lizards", package = "brglm2")
 links <- lapply(c("logit", "probit", "cloglog", "cauchit"), make.link)
 
-tol <- 1e-10
+tol <- 1e-08
 for (l in seq_along(links)) {
     expect_warning(
         lizardsBRlegacy <- brglm(cbind(grahami, opalinus) ~ height + diameter +
diff --git a/inst/tinytest/test-bracl.R b/inst/tinytest/test-bracl.R
index 936cf74..7712a96 100644
--- a/inst/tinytest/test-bracl.R
+++ b/inst/tinytest/test-bracl.R
@@ -85,8 +85,8 @@ s2p <- summary(fit_vgam_p)
 
 tol <- 1e-06
 ## summary method for bracl returns the correct coef mat
-expect_equal(coef(s1), coef(s2), tolerance = tol, check.attributes = FALSE)
-expect_equal(coef(s1p), coef(s2p), tolerance = tol, check.attributes = FALSE)
+expect_equal(coef(s1), s2@coef3, tolerance = tol, check.attributes = FALSE)
+expect_equal(coef(s1p), s2p@coef3, tolerance = tol, check.attributes = FALSE)
 
 newdata <- expand.grid(gender = c("male", "female"),  religion = c("moderate", "fundamentalist"))
 ## predict.bracl works as expected
@@ -94,10 +94,10 @@ pp <- predict(fit_bracl_p, newdata = stemcell, type = "probs")
 p <- predict(fit_bracl, newdata = stemcell, type = "probs")
 expect_equal(predict(fit_vgam_p, type = "response"),
              pp[19:24, ],
-             tolerance = 1e-08, check.attributes = FALSE)
+             tolerance = 1e-06, check.attributes = FALSE)
 expect_equal(predict(fit_vgam, type = "response"),
              p[19:24, ],
-             tolerance = 1e-08, check.attributes = FALSE)
+             tolerance = 1e-06, check.attributes = FALSE)
 
 ## no intercept returns warning
 expect_warning(
diff --git a/inst/tinytest/test-jeffreys.R b/inst/tinytest/test-jeffreys.R
index 8faeae7..7205c95 100644
--- a/inst/tinytest/test-jeffreys.R
+++ b/inst/tinytest/test-jeffreys.R
@@ -41,7 +41,7 @@ data("lizards", package = "brglm2")
 links <- lapply(c("logit", "probit", "cloglog", "cauchit"), make.link)
 
 
-tol <- 1e-10
+tol <- 1e-08
 for (l in seq_along(links)) {
     expect_warning(
         lizardsBRlegacy <- brglm(cbind(grahami, opalinus) ~ height + diameter +
diff --git a/inst/tinytest/test-transformation.R b/inst/tinytest/test-transformation.R
index 096c0f1..b5cff7a 100644
--- a/inst/tinytest/test-transformation.R
+++ b/inst/tinytest/test-transformation.R
@@ -24,7 +24,7 @@ c_sqrt[5] <- c_sqrt[5]^2
 c_inverse <- coef(mod_inverse, model = "full")
 c_inverse[5] <- 1/c_inverse[5]
 
-tol <- sqrt(.Machine$double.eps)
+tol <- 1e-08
 ## ML estimate of gamma dispersion from brglmFit is invariant to trasnformation
 expect_equal(c_identity, c_log, tolerance = tol, check.attributes = FALSE)
 expect_equal(c_identity, c_sqrt, tolerance = tol, check.attributes = FALSE)
diff --git a/man/brglm2.Rd b/man/brglm2.Rd
index bf2f5cb..3f7191f 100644
--- a/man/brglm2.Rd
+++ b/man/brglm2.Rd
@@ -3,6 +3,7 @@
 \docType{package}
 \name{brglm2}
 \alias{brglm2}
+\alias{brglm2-package}
 \title{brglm2: Bias Reduction in Generalized Linear Models}
 \description{
 Estimation and inference from generalized linear models using

From dfd9c38fdf90f5652f316550a9a7a00a856427d0 Mon Sep 17 00:00:00 2001
From: Ioannis Kosmidis <ioannis.kosmidis@warwick.ac.uk>
Date: Thu, 5 Oct 2023 12:23:33 +0100
Subject: [PATCH 4/7] Vignette and example updates

---
 .Rbuildignore                  | 2 ++
 .gitignore                     | 4 +++-
 NEWS.md                        | 1 +
 R/brglmFit.R                   | 1 +
 R/brmultinom.R                 | 2 +-
 man/brmultinom.Rd              | 2 +-
 vignettes/adjacent.Rmd         | 2 +-
 vignettes/iteration.Rmd        | 8 ++++----
 vignettes/multinomial.Rmd      | 4 ++--
 vignettes/negativeBinomial.Rmd | 2 +-
 10 files changed, 17 insertions(+), 11 deletions(-)

diff --git a/.Rbuildignore b/.Rbuildignore
index 341645f..4449f79 100644
--- a/.Rbuildignore
+++ b/.Rbuildignore
@@ -30,3 +30,5 @@ README_cache
 ^CRAN-RELEASE$
 ^code_of_conduct\.md$
 ^CRAN-SUBMISSION$
+^doc$
+^Meta$
diff --git a/.gitignore b/.gitignore
index 06d1b0a..b4e86b5 100644
--- a/.gitignore
+++ b/.gitignore
@@ -115,4 +115,6 @@ flycheck_*.el
 .dir-locals.el
 
 # network security
-/network-security.data
\ No newline at end of file
+/network-security.data
+/doc/
+/Meta/
diff --git a/NEWS.md b/NEWS.md
index 31e9e89..785a2e2 100644
--- a/NEWS.md
+++ b/NEWS.md
@@ -6,6 +6,7 @@
 
 * `brglm_control()` now checks that the supplied value of `max_step_factor` is numeric and greater or equal to `1`. If not, then it is set to the default value of `12`.
 
+* Vignette updates
 
 
 # brglm2 0.9.1
diff --git a/R/brglmFit.R b/R/brglmFit.R
index 3ab1284..d669929 100644
--- a/R/brglmFit.R
+++ b/R/brglmFit.R
@@ -992,6 +992,7 @@ brglmFit <- function(x, y, weights = rep(1, nobs), start = NULL, etastart = NULL
         ## residuals and working_weights
 
         quantities <- key_quantities(c(betas, dispersion), y = y, level = 2 * !no_dispersion, scale_totals = has_fixed_totals, qr = TRUE)
+
         qr.Wx <- quantities$qr_decomposition
 
         mus <- quantities$mus
diff --git a/R/brmultinom.R b/R/brmultinom.R
index 5a5aeff..5d1fd6a 100644
--- a/R/brmultinom.R
+++ b/R/brmultinom.R
@@ -142,7 +142,7 @@
 #' contrasts(hepat$type) <- contr.treatment(3, base = 1)
 #'
 #' # Maximum likelihood estimation fails to converge because some estimates are infinite
-#' hepML <- brmultinom(type ~ group * time, data = hepat, weights = counts, type = "ML")
+#' hepML <- brmultinom(type ~ group * time, data = hepat, weights = counts, type = "ML", slowit = 0.1)
 #'
 #' # Mean bias reduction returns finite estimates
 #' hep_meanBR <- brmultinom(type ~ group * time, data = hepat, weights = counts, type = "AS_mean")
diff --git a/man/brmultinom.Rd b/man/brmultinom.Rd
index 4894abd..4c1959b 100644
--- a/man/brmultinom.Rd
+++ b/man/brmultinom.Rd
@@ -160,7 +160,7 @@ hepat$type <- factor(hepat$type, labels = c("noDisease", "C", "nonABC"))
 contrasts(hepat$type) <- contr.treatment(3, base = 1)
 
 # Maximum likelihood estimation fails to converge because some estimates are infinite
-hepML <- brmultinom(type ~ group * time, data = hepat, weights = counts, type = "ML")
+hepML <- brmultinom(type ~ group * time, data = hepat, weights = counts, type = "ML", slowit = 0.1)
 
 # Mean bias reduction returns finite estimates
 hep_meanBR <- brmultinom(type ~ group * time, data = hepat, weights = counts, type = "AS_mean")
diff --git a/vignettes/adjacent.Rmd b/vignettes/adjacent.Rmd
index 1d43490..3db26c4 100644
--- a/vignettes/adjacent.Rmd
+++ b/vignettes/adjacent.Rmd
@@ -7,7 +7,7 @@ bibliography: brglm2.bib
 nocite: |
   @kosmidis:2019
 vignette: >
-  %\VignetteIndexEntry{Adjacent category logit models using brglm2}
+  %\VignetteIndexEntry{Adjacent category logit models using **brglm2**}
   %\VignetteEngine{knitr::rmarkdown}
   %\VignetteEncoding{UTF-8}
 ---
diff --git a/vignettes/iteration.Rmd b/vignettes/iteration.Rmd
index c243373..1b9a39c 100755
--- a/vignettes/iteration.Rmd
+++ b/vignettes/iteration.Rmd
@@ -185,7 +185,7 @@ where $\rho = (\rho_1, \ldots, \rho_n)^T$ with $\rho_i = h_i \{2 d_i'/(d_i w_i)
 + $s_\beta$, $i_{\beta\beta}$, $A_\beta$
 + $s_\phi$, $i_{\phi\phi}$, $A_\phi$
 + Starting values $\beta^{(0)}$ and $\phi^{(0)}$
-+ $\epsilon > 0$: tolerance for the $L1$ norm of the direction before reporting convergence
++ $\epsilon > 0$: tolerance for the $L^\infty$ norm of the search direction before reporting convergence
 + $M$: maximum number of halving steps that can be taken
 
 ### Output
@@ -214,7 +214,7 @@ where $\rho = (\rho_1, \ldots, \rho_n)^T$ with $\rho_i = h_i \{2 d_i'/(d_i w_i)
 
 8. $v_\phi^{(m)} \leftarrow \upsilon_\phi^{(k)}$
 
-9. $d \leftarrow \left|v_\beta^{(m)}\right|_1 + \left|v_\phi^{(m)}\right|$
+9. $d \leftarrow \left|\left|(v_\beta^{(m)}, v_\phi^{(m)})\right|\right|_\infty$
 
 *Update parameters*
 
@@ -230,7 +230,7 @@ where $\rho = (\rho_1, \ldots, \rho_n)^T$ with $\rho_i = h_i \{2 d_i'/(d_i w_i)
 
 *Continue or break halving within inner loop*
 
-14. if $m + 1 < M$ and $\left|v_\beta^{(m + 1)}\right|_1 + \left|v_\phi^{(m + 1)}\right| > d$
+14. if $m + 1 < M$ and $\left|\left|(v_\beta^{(m + 1)}, v_\phi^{(m + 1)})\right|\right|_\infty > d$
 
 	14.1. $m \leftarrow m + 1$
 
@@ -248,7 +248,7 @@ where $\rho = (\rho_1, \ldots, \rho_n)^T$ with $\rho_i = h_i \{2 d_i'/(d_i w_i)
 
 *Continue or break outer loop*
 
-16. if $k + 1 < K$ and $\left|\upsilon_\beta^{(k + 1)}\right|_1 + \left|\upsilon_\phi^{(k + 1)}\right| > \epsilon$
+16. if $k + 1 < K$ and $\left|\left|(\upsilon_\beta^{(k + 1)}, \upsilon_\phi^{(k + 1)})\right|\right|_\infty > \epsilon$
 
 	16.1 $k \leftarrow k + 1$
 
diff --git a/vignettes/multinomial.Rmd b/vignettes/multinomial.Rmd
index 290f37d..28fb283 100644
--- a/vignettes/multinomial.Rmd
+++ b/vignettes/multinomial.Rmd
@@ -73,12 +73,12 @@ particular model and data combination.
 
 Let's scale the frequencies in `alligators` by 3 in order to get a sparser data set. The differences between maximum likelihood and mean and median bias reduction should be more apparent on the resulting data set. Here we have to "slow-down" the Fisher scoring iteration (by scaling the step-size), because otherwise the Fisher information matrix quickly gets numerically rank-deficient. The reason is data separation [@albert:84].
 ```{r, echo = TRUE, error = TRUE}
-all_ml_sparse <- update(all_ml, weights = round(freq/3), slowit = 0.2)
+all_ml_sparse <- update(all_ml, weights = round(freq/3), slowit = 0.1)
 summary(all_ml_sparse)
 ```
 Specifically, judging from the estimated standard errors, the estimates for `(Intercept)`, `lakeHancock`, `lakeOklawaha` and `lakeTrafford` for `Reptile` and `lakeHancock` for `Bird` seem to be infinite.
 
-To quickly check if that's indeed the case we can use the `check_infinite_estimates()` method of the [**detectseparation**][https://cran.r-project.org/package=detectseparation] R package.
+To quickly check if that's indeed the case we can use the `check_infinite_estimates()` method of the [**detectseparation**](https://cran.r-project.org/package=detectseparation) R package.
 ```{r, echo = TRUE}
 library("detectseparation")
 se_ratios <- check_infinite_estimates(all_ml_sparse)
diff --git a/vignettes/negativeBinomial.Rmd b/vignettes/negativeBinomial.Rmd
index 669b4bf..a225061 100644
--- a/vignettes/negativeBinomial.Rmd
+++ b/vignettes/negativeBinomial.Rmd
@@ -5,7 +5,7 @@ date: "12 June 2021"
 output: rmarkdown::html_vignette
 bibliography: brglm2.bib
 vignette: >
-  %\VignetteIndexEntry{Negative Binomial regression using **brglm2**}
+  %\VignetteIndexEntry{Negative binomial regression using **brglm2**}
   %\VignetteEngine{knitr::rmarkdown}
   %\VignetteEncoding{UTF-8}
 ---

From 37236c819720ab47bda318bc469d6358e68d9c77 Mon Sep 17 00:00:00 2001
From: Ioannis Kosmidis <ioannis.kosmidis@warwick.ac.uk>
Date: Thu, 5 Oct 2023 12:41:54 +0100
Subject: [PATCH 5/7] Version bump

---
 DESCRIPTION | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/DESCRIPTION b/DESCRIPTION
index 2522b27..abee498 100644
--- a/DESCRIPTION
+++ b/DESCRIPTION
@@ -1,6 +1,6 @@
 Package: brglm2
 Title: Bias Reduction in Generalized Linear Models
-Version: 0.9.999
+Version: 0.9.2
 Authors@R: c(person(given = "Ioannis", family = "Kosmidis", role = c("aut", "cre"), email = "ioannis.kosmidis@warwick.ac.uk", comment = c(ORCID = "0000-0003-1556-0302")),
 	     person(given = c("Euloge", "Clovis"), family = c("Kenne Pagui"), role = "aut", email = "kenne@stat.unipd.it"),
 	     person(given = "Kjell", family = "Konis", role = "ctb", email = "kjell.konis@me.com"),

From cc80a2f792e05c6e09921fca9713e72234f6ae82 Mon Sep 17 00:00:00 2001
From: Ioannis Kosmidis <ioannis.kosmidis@warwick.ac.uk>
Date: Wed, 11 Oct 2023 16:17:52 +0100
Subject: [PATCH 6/7] Attempt to CRAN

---
 CRAN-SUBMISSION | 3 +++
 1 file changed, 3 insertions(+)
 create mode 100644 CRAN-SUBMISSION

diff --git a/CRAN-SUBMISSION b/CRAN-SUBMISSION
new file mode 100644
index 0000000..9aa33bf
--- /dev/null
+++ b/CRAN-SUBMISSION
@@ -0,0 +1,3 @@
+Version: 0.9.2
+Date: 2023-10-11 15:17:41 UTC
+SHA: 37236c819720ab47bda318bc469d6358e68d9c77

From 2769231f693b383d6c394a9138f11b2bb25877f1 Mon Sep 17 00:00:00 2001
From: Ioannis Kosmidis <ioannis.kosmidis@warwick.ac.uk>
Date: Thu, 12 Sep 2024 18:55:37 +0100
Subject: [PATCH 7/7] deleted temporary files

---
 CRAN-SUBMISSION | 3 ---
 1 file changed, 3 deletions(-)
 delete mode 100644 CRAN-SUBMISSION

diff --git a/CRAN-SUBMISSION b/CRAN-SUBMISSION
deleted file mode 100644
index 9aa33bf..0000000
--- a/CRAN-SUBMISSION
+++ /dev/null
@@ -1,3 +0,0 @@
-Version: 0.9.2
-Date: 2023-10-11 15:17:41 UTC
-SHA: 37236c819720ab47bda318bc469d6358e68d9c77