Jump to content


Photo

Question of output for individual modelling result


  • Please log in to reply
6 replies to this topic

#1 Emily H

Emily H

    Newbie

  • Members
  • Pip
  • 3 posts

Posted 08 August 2022 - 01:09 PM

Dear all, 

 

I would like to confirm the following issue:

 

1. How did the IWRES be calculated from the output data? Is the IWRES=(DV-predicted)/weight ?

 

2. what is detail calculation information of LOESS regression including (1) degree of local polynomials, (2) smoothing parameter, and (3) weight function in the IWRES vs IPRED (or IWRES vs IVAR ) 

 

Thanks in Advance

 

Emily



#2 smouksassi1

smouksassi1

    Advanced Member

  • Members
  • PipPipPip
  • 231 posts
  • LocationMontreal

Posted 08 August 2022 - 01:18 PM

see equation 16 in my paper for IWRES definition this is a standard residual used in the pmx work 
https://ascpt.online...1002/psp4.12161

the developers can comment on what is being used in the Phoenix software as for the loess you cannot control anything about it except the n of points.

developers can disclose what is being used under the hood ?



#3 bwendt@certara.com

bwendt@certara.com

    Advanced Member

  • Administrators
  • 282 posts

Posted 09 August 2022 - 08:13 AM

Hi Emily,

 

the Phoenix Loess function uses the Cleveland-Grosse  Bell labs fortran 77 implementation in Netlib.  See http://www.netlib.org/a/loess for a description.  The smoothing parameter, alpha, is determined automatically and depends on the number of points:          if(n.gt.10) alpha=1         if(n.gt.10) alpha=.9         if(n.gt.15) alpha=.8         if(n.gt.20) alpha=.7         if(n.gt.50) alpha=.6         if(n.ge.100) alpha=.5         if(n.ge.200) alpha=.5         if(n.ge.500) alpha=.5         if(n.ge.1000) alpha=0.5  In most of the case, the default smoothing value is good enough.

 

Bernd



#4 smouksassi1

smouksassi1

    Advanced Member

  • Members
  • PipPipPip
  • 231 posts
  • LocationMontreal

Posted 09 August 2022 - 11:50 AM

in the future version of Phoenix you will be able to have more control on thow the smoothing alpha is set



#5 Emily H

Emily H

    Newbie

  • Members
  • Pip
  • 3 posts

Posted 10 August 2022 - 06:18 AM

Dear all,

 

Thanks for all of your reply.  Based on the article and link, I also have some questions: 

 

1. IWRES- I would like to confirm If the weight for an additive error model under individual mode is the square root of the variance of the residual error? cause  we are NOT using population PK model but individual mode.

 

2. The weight  of  LOESS created from Phoenix is  described  in the code  written in Fortran. However, I am not able to get understand with it. Could anyone help me to transfer into  a layman’s term how the weight is calculated ?  

 

Thank you very much

 

Emily



#6 smouksassi1

smouksassi1

    Advanced Member

  • Members
  • PipPipPip
  • 231 posts
  • LocationMontreal

Posted 10 August 2022 - 07:58 AM

https://onlinehelp.c...=IWRES&rhsyns=

 

here you have the online help definition of IWRES:

WRES=(DV – IPRED)/STD (1)

(see also the definition for STD for additive and other error models)

 

the loess "span" depends on the n of neighbor points not sure I understand the "Weight" of loess can you clarify ?



#7 Emily H

Emily H

    Newbie

  • Members
  • Pip
  • 3 posts

Posted 11 August 2022 - 12:36 PM

Dear sir,

 

For the weight function in LOESS, it is to give individual observations in the sum of squared residuals a different weight in the local fitting criterion.

 

thank you very much 

 

Emily






0 user(s) are reading this topic

0 members, 0 guests, 0 anonymous users