search

andrewhooker / PopED

Population Experimental Design (PopED) in R
https://andrewhooker.github.io/PopED/
GNU Lesser General Public License v3.0
33 stars 21 forks source link

Simplify derivatives #34

Closed martin-gmx closed 6 years ago

martin-gmx commented 6 years ago

Hi Andy. These changes should allow us to more easily add other forms of derivatives.

II am still trying to get my head around this double derivatives - that we take the derivative of v wrt all parameters, but v already linearizes (i.e., takes the derivative of all parameters). Couldn't we try to use the chain rule and remove this multiplication of derivatives?

Cheers Martin

mattfidler commented 6 years ago

Hi @andrewhooker and @martin-gmx ; Does this imply you need to use deSolve; I was working something out with nlmixr that integrates with PopED, but uses RxODE. I haven't tested it but will abandon it if you require deSolve

Sibojang9 commented 6 years ago

My personal experience is that RxODE is much easier to use than deSolve. I would vote for PopED+RxODE.

andrewhooker commented 6 years ago

Hi

No. We do not require deSolve. Other tools and packages can be used. All of the standard model simulation package for population models can be used as input to poped.

Andy On 31 Oct 2018, 19:34 +0100, Sibo Jiang notifications@github.com, wrote:

My personal experience is that RxODE is much easier to use than deSolve. I would vote for PopED+RxODE. — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub, or mute the thread.

martin-gmx commented 6 years ago

The user provides the model solution - and in the future we will (hopefully) be able to use also the partial derivatives of the solution w.r.t. the model parameters, if the user provides them.

mattfidler commented 6 years ago

You can request derivatives from RxODE too, perhaps this couod be helpful here Martin.

On Sat, Nov 3, 2018, 11:36 AM martin-gmx notifications@github.com wrote:

The use provides the model solution - and in the future we will (hopefully) be able to use also the partial derivatives of the solution w.r.t. the model parameters, if the user provides them.

— You are receiving this because you commented.

Reply to this email directly, view it on GitHub https://github.com/andrewhooker/PopED/pull/34#issuecomment-435601845, or mute the thread https://github.com/notifications/unsubscribe-auth/AAfa2tskxAtWF3kl5OiwberjCuU4ZN-0ks5urcX8gaJpZM4Xy6pa .

andrewhooker commented 6 years ago

Looks good Martin!