plot_etiology_regression: visualize the etiology regression with a continuous covariate
In oslerinhealth/baker: Bayesian Analysis Kit for Etiology Research

Description Usage Arguments Value References See Also Examples

View source: R/plot-etiology-regression.R

This function visualizes the etiology regression against one continuous covariate, e.g., enrollment date. (NB: dealing with NoA, multiple-pathogen causes, other continuous covariates? also there this function only plots the first slice - so generalization may be useful - give users an option to choose slice s; currently default to the first slice.)

plot_etiology_regression(
  DIR_NPLCM,
  stratum_bool,
  slice = 1,
  plot_basis = FALSE,
  truth = NULL,
  RES_NPLCM = NULL,
  do_plot = TRUE,
  do_rug = TRUE,
  return_metric = TRUE,
  plot_ma_dots = FALSE
)

`DIR_NPLCM`	File path to the folder containing posterior samples
`stratum_bool`	a vector of TRUE/FALSE with TRUE indicating the rows of subjects to include
`slice`	integer; specifies which slice of bronze-standard data to visualize; Default to 1.
`plot_basis`	TRUE for plotting basis functions; Default to FALSE
`truth`	a list of truths computed from true parameters in simulations; elements: Eti, FPR, PR_case,TPR; All default to `NULL` in real data analyses. Currently only works for one slice of bronze-standard measurements (in a non-nested model). Eti matrix of # of rows = # of subjects, # columns: `length(cause_list)` for Eti FPR matrix of # of rows = # of subjects, # columns: `ncol(data_nplcm$Mobs$MBS$MBS1)` PR_case matrix of # of rows = # of subjects, # columns: `ncol(data_nplcm$Mobs$MBS$MBS1)` TPR a vector of length identical to `PR_case`
`RES_NPLCM`	pre-read res_nplcm; default to NULL.
`do_plot`	TRUE for plotting
`do_rug`	TRUE for plotting
`return_metric`	TRUE for showing overall mean etiology, quantiles, s.d., and if `truth$Eti` is supplied, coverage, bias, truth and integrated mean squared errors (IMSE).
`plot_ma_dots`	plot moving averages among case and controls if TRUE; Default to FALSE.

A figure of etiology regression curves and some marginal positive rate assessment of model fit; See example for the legends.

See example figures

A Figure using simulated data for six pathogens: https://bit.ly/2FWoYeM
The legends for the figure above: https://bit.ly/2OU8F60

Other visualization functions: plot_BrS_panel(), plot_SS_panel(), plot_check_common_pattern(), plot_check_pairwise_SLORD(), plot_etiology_side_by_side(), plot_etiology_strat_nested(), plot_etiology_strat(), plot_group_etiology(), plot_panels(), plot_pie_panel(), plot_selected_etiology(), plot_subwt_regression()

## Not run: 
# legend.text = c("[UPPER FIGURES]",
                "observed prevalence: cases",
                "observed prevalence: controls",
                "fitted prevalence: cases",
                "fitted prevalence: controls",
                "true positive rate: mean",
                "true positive rate: 95%CI",
                "[BOTTOM FIGURES]",
                "etiology curve: mean",
                "overall etiology: mean",
                "overall etiology: 95%CI","","","")
legend.col=c("white","black","dodgerblue2","black","dodgerblue2","red","red",
             "white","springgreen4","orange","orange","white","white","white")
legend.lty=c(1,2,2,1,1,1,2,1,1,1,2)
legend.lwd=c(2,2,2,2,2,2,2,2,2,2,2,2,2,2)
legend("topleft",legend=legend.text,
       lty=legend.lty,lwd=legend.lwd,
       col=legend.col,ncol=2,
       y.intersp=1.5,cex=1.6,box.col=NA)
       
## End(Not run)