60 auc estimation streamline conc scale for estimation add argument #64
Conversation
…-scale, updated the function logic and documentation, added re-parameterized functions with log10 scale x
gracezhihuizhao
requested review from
kpaulfriedman,
brown-jason and
sedavid01
Merge branch 'dev' of https://github.com/USEPA/CompTox-ToxCast-tcplFit2 into 60-auc-estimation-streamline-conc-scale-for-estimation-add-argument # Conflicts: # R/concRespCore.R
…the plots; Updated the bi-phasic curve section after the poly2 update (more development needed).
| logpoly1 <- function(ps, x) { | ||
| return(ps[1]*(10^x)) | ||
| } | ||
|
|
There was a problem hiding this comment.
- poly 2 should be moved up here with poly1
- also, suggest moving the power with the poly2 - i.e. poly models followed by power followed by exponential models
R/get_AUC.R
Outdated
| @@ -19,6 +19,9 @@ | |||
| #' @param ps Numeric vector (or list) of model parameters for the specified model in `fit_method`. | |||
| #' @param return.abs Logical argument, defaults to FALSE. | |||
There was a problem hiding this comment.
@gracezhihuizhao, can you please provide a little more detail with regard to what this logical argument is meant to do? I assume this returns the absolute value of the AUC. However, current documentation is very basic and uninformative.
vignettes/get_AUC_demo.Rmd
Outdated
There was a problem hiding this comment.
@madison-feshuk and @brown-jason, this vignette may need to be double checked for any additional explanatory text and/or polishing before merging into 'dev' or release of the next package version.
vignettes/get_AUC_demo.Rmd
Outdated
There was a problem hiding this comment.
@gracezhihuizhao, please check log-scale plots (particularly Fig 1 and Fig 3). Your fix for this may be in another branch , but please check.
There was a problem hiding this comment.
Yes, the plots have the same issue you identified in PR #87. The fix in that PR will also fix the Fig 1 and Fig 3.
There was a problem hiding this comment.
@gracezhihuizhao, in code documentation of the re-param functions to log-scale it may be helpful to have a code comment above the function with the respective 'normal'/'raw' scale parameterized version of the function equation for easy reference.
…ument to post_hit_AUC as well. Added normal parameterization references for log-scale re-parameterized functions.
…stimation-add-argument
Things that are nice to have from the ticket #60 :
apply(), in my opinion it will make things easier if the function only return a single value.