ASMODEE is an outbreak detection algorithm, implemented as asmodee in the package trendbreaker. Here it is evaluated on simulations (time series generated by the function project of the package projections) and on the English CCG's (see https://covid19-nhs-pathways-asmodee.netlify.app/).
The method and results are presented in a preprint (first version here, revised version accepted for publication).
With the ASMODEE method evaluate_aic, the only randomness in this analysis comes from the simulations. Since random seeds are not set, each run will produce slightly different results. However all relevant intermediate objects are stored in data/: if you want to reproduce exactly the results presented in the upcoming publication, you can set compute_simulations <- FALSE at the beginning of the script, the simulations will then be loaded from data/simulations.rds.
It takes about 4 hours to run on a laptop for model selection through AIC (method = evaluate_aic) and about 2 days with cross-validation (method = evaluate_resampling).
In this repository further results, not included in the article, can be found in img/sim-project/ in the subfolders scenario-*. We caution however that, when not computed on either class decrease or increase, those are not directly comparable with the results shown in the article. Here the scores (sensitivity, specificity, precision, ...) are defined either separately on different periods or reflect a multiclass classification framing. Further plots about the CCG's can be found in img/ccgs/.
The packages trending and trendbreaker where downloaded on 14 February 2021 from their respective "bootstrap" branches via:
remotes::install_github("reconhub/trending@bootstrap")
remotes::install_github("reconhub/trendbreaker@bootstrap")
Overall session information:
R version 4.0.3 (2020-10-10)
Platform: x86_64-apple-darwin17.0 (64-bit)
Running under: macOS Big Sur 10.16
Matrix products: default
LAPACK: /Library/Frameworks/R.framework/Versions/4.0/Resources/lib/libRlapack.dylib
locale:
[1] en_US.UTF-8/en_US.UTF-8/en_US.UTF-8/C/en_US.UTF-8/en_US.UTF-8
attached base packages:
[1] tools grid stats graphics grDevices utils datasets methods base
other attached packages:
[1] dplyr_1.0.4 dtplyr_1.0.1 data.table_1.13.6 surveillance_1.18.0
[5] xtable_1.8-4 sp_1.4-4 ggtext_0.1.0 gridExtra_2.3
[9] RColorBrewer_1.1-2 trendbreaker_0.0.2 caret_6.0-86 ggplot2_3.3.3
[13] lattice_0.20-41 projections_0.5.1 tidyr_1.1.2 distcrete_1.0.3
[17] epitrix_0.2.2 MASS_7.3-53 here_1.0.0
loaded via a namespace (and not attached):
[1] Rcpp_1.0.6 lubridate_1.7.9.2 deldir_0.2-3 class_7.3-17
[5] assertthat_0.2.1 rprojroot_2.0.2 ipred_0.9-9 foreach_1.5.1
[9] R6_2.5.0 plyr_1.8.6 stats4_4.0.3 tensor_1.5
[13] pillar_1.4.7 rlang_0.4.10 goftest_1.2-2 rpart_4.1-15
[17] Matrix_1.2-18 splines_4.0.3 incidence_1.7.3 gower_0.2.2
[21] stringr_1.4.0 polyclip_1.10-0 munsell_0.5.0 gridtext_0.1.1
[25] spatstat.data_1.5-2 compiler_4.0.3 pkgconfig_2.0.3 trendeval_0.0.1
[29] mgcv_1.8-33 nnet_7.3-14 tidyselect_1.1.0 tibble_3.0.6
[33] prodlim_2019.11.13 codetools_0.2-18 crayon_1.4.1 withr_2.4.1
[37] recipes_0.1.15 ModelMetrics_1.2.2.2 nlme_3.1-150 gtable_0.3.0
[41] lifecycle_0.2.0 DBI_1.1.0 polyCub_0.7.1 magrittr_2.0.1
[45] pROC_1.17.0.1 scales_1.1.1 stringi_1.5.3 reshape2_1.4.4
[49] spatstat_1.64-1 timeDate_3043.102 xml2_1.3.2 spatstat.utils_1.17-0
[53] ellipsis_0.3.1 generics_0.1.0 vctrs_0.3.6 lava_1.6.8.1
[57] iterators_1.0.13 glue_1.4.2 purrr_0.3.4 trending_0.0.2.9
[61] abind_1.4-5 survival_3.2-7 colorspace_2.0-0
Code, data and images are made available under a CC0 licence, i.e. others may freely build upon, enhance and reuse the works for any purposes without restriction under copyright or database law.