Default initialRidgeVariance for iterative hard thresholding too large

[MathildeTans]

Describe the bug The default value for the variable 'initialRidgeVariance' (part of iterative hard thresholding) is too large and can cause inability to converge if it is not set properly. I have tried to commit the required edit, but I did not have access.

Set up (please run in R "sessionInfo()" and copy the output here): R version 4.0.3 (2020-10-10) Platform: x86_64-w64-mingw32/x64 (64-bit) Running under: Windows 10 x64 (build 22000)

Matrix products: default

locale: [1] LC_COLLATE=English_United States.1252 LC_CTYPE=English_United States.1252 LC_MONETARY=English_United States.1252 LC_NUMERIC=C
[5] LC_TIME=English_United States.1252

attached base packages: [1] stats graphics grDevices utils datasets methods base

loaded via a namespace (and not attached): [1] Rcpp_1.0.6 Andromeda_0.4.0 rstudioapi_0.13 DatabaseConnector_3.0.0 magrittr_2.0.1 tidyselect_1.1.0
[7] bit_4.0.4 FeatureExtraction_3.1.0 R6_2.5.0 rlang_0.4.10 blob_1.2.1 dplyr_1.0.3
[13] tools_4.0.3 DBI_1.1.1 SqlRender_1.6.8 ellipsis_0.3.1 assertthat_0.2.1 bit64_4.0.5
[19] digest_0.6.27 tibble_3.0.5 lifecycle_0.2.0 crayon_1.3.4 rJava_0.9-13 purrr_0.3.4
[25] ParallelLogger_3.0.0 vctrs_0.3.6 zeallot_0.1.0 memoise_1.1.0 glue_1.4.2 RSQLite_2.2.2
[31] PatientLevelPrediction_4.3.10 compiler_4.0.3 pillar_1.4.7 generics_0.1.0 jsonlite_1.7.2 pkgconfig_2.0.3

To Reproduce In CyclopsSettins.R, function 'setIterativeHardThresholding', the initialRidgeVariance has default value 1000. I advise changing this to 0.1. This is on line 190.

PLP Log File N/A

Additional context I should note that I deduced that it should be 0.1 instead of 1000 using just one PLP model. It might be different for other problems. I would recommend changing it since an initialRidgeVariance of 1000 is likely to be too large. But I would also keep in mind that there might be a better default value for this variable.

[jreps]

Thanks for figuring this out! To make the edit you can fork PLP (develop branch) to your own repositories and then make a pull request into develop from your fork. I'll mark this as a good first commit.