VPC for log TBS data

[PerOG7]

Hi Developers,

Is there a way for doing VPCs on models with log transform-both-sides residual error models, so that the non-transformed data is plotted, i.e. consistent with the lnDV=1 flag in PsN?

Kind Regards,
Per

[bguiastr]

Hello,

There is currently no built-in feature to transform the data in VPCs as in PsN lnDV=x.

You could try to point xpose to the original DV (i.e., non-log transformed) rather than the log transformed one? (note: this assumes that your original DV is in your observation and simulation tables).

xpdb <- set_var_types(xpdb, dv = "ODV")

If you don't have the original DV in the output tables you could transform it using:

xpdb <- xpdb %>% mutate(ODV = exp(DV)) %>% set_var_types(dv = "ODV")

And then run the vpc on that transformed dataset.

Best of luck!

[PerOG7]

Thanks! Per

…

On Tue, 24 May 2022, 07:57 Benjamin, ***@***.***> wrote: Hello, There is currently no built-in feature to transform the data in VPCs as in PsN lnDV=x. You could try to point xpose to the original DV (i.e., non-log transformed) rather than the log transformed one? (note: this assumes that your original DV is in your observation and simulation tables). xpdb <- set_var_types(xpdb, dv = "ODV") If you don't have the original DV in the output tables you could transform it using: xpdb <- xpdb %>% mutate(ODV = exp(DV)) %>% set_var_types(dv = "ODV") And then run the vpc on that transformed dataset. Best of luck! — Reply to this email directly, view it on GitHub <#201 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AYA3HJVZ3TNYOCSAROGVSRDVLR4VRANCNFSM5WWJM4UA> . You are receiving this because you authored the thread.Message ID: ***@***.***>