Fix pareto warnings

DESCRIPTION

@@ -86,5 +86,5 @@ Additional_repositories:
     https://mc-stan.org/r-packages/
 LazyData: TRUE
 Roxygen: list(markdown = TRUE)
-RoxygenNote: 7.2.3
+RoxygenNote: 7.3.1
 VignetteBuilder: knitr, rmarkdown

NEWS.md

@@ -4,6 +4,9 @@ If you read this from a place other than <https://mc-stan.org/projpred/news/inde
 
 # projpred 2.8.0.9000
 
+## Minor changes
+
+* Use the updated threshold for high Pareto-$\hat{k}$ values presented by Vehtari et al. (2024, "Pareto smoothed importance sampling", *Journal of Machine Learning Research*, 25(72):1-58, <https://www.jmlr.org/papers/v25/19-556.html>). This threshold depends on the Monte Carlo sample size and is often close to the former fixed threshold of 0.7 (a short introduction may also be found in the [LOO glossary](https://mc-stan.org/loo/reference/loo-glossary.html)). Correspondingly, the former "secondary" threshold of 0.5 is not used anymore either.
 
 # projpred 2.8.0
 

R/varsel.R

@@ -5,7 +5,7 @@
 #' known as solution path), i.e., the best submodel for each submodel size
 #' (number of predictor terms). The evaluation part determines the predictive
 #' performance of the submodels along the predictor ranking. A special method is
-#' [varsel.vsel()] because it re-uses the search results from an earlier
+#' [varsel.vsel()] which re-uses the search results from an earlier
 #' [varsel()] (or [cv_varsel()]) run, as illustrated in the main vignette.
 #'
 #' @param object An object of class `refmodel` (returned by [get_refmodel()] or
@@ -141,8 +141,9 @@
 #'
 #'   L1 search is faster than forward search, but forward search may be more
 #'   accurate. Furthermore, forward search may find a sparser model with
-#'   comparable performance to that found by L1 search, but it may also start
-#'   overfitting when more predictors are added.
+#'   comparable performance to that found by L1 search, but it may also
+#'   overfit when more predictors are added. This overfit can be detected
+#'   by running search validation (see [cv_varsel()]).
 #'
 #'   An L1 search may select an interaction term before all involved lower-order
 #'   interaction terms (including main-effect terms) have been selected. In