Verhulst Curve Fit

[azikorus-disabl]

Collab,

We recently performed an experiment which is sort of the opposite of Verhulst. A special cell culture, suspended in an appropriate media, was held at a constant temperature.

The study was to analyze the rate at which cell death occurred at different temperatures.

For this study, the cell population's viability was sampled at 6 minute intervals up to 2 hours.

The resulting curve appears somewhat like an inverted Verhulst curve.

Using a model from an earlier post by Jean showing the Verhulst model, I tweaked the coefficients to get a best visual match.

Apparently, Verhulst is not the best model to fit my data.

Does Jean or any Collab have a suggestion for a better fit model?

Thanks,

Art

[TomGutman]

I rarely worry about what Jean and Theodore are pushing.

In general, ODEs do not have analytic solutions. There are some exceptions, for particularly simple cases, but mostly you have to rely on good numeric solvers. Basically the ODE is the solution. Excel is a great spread sheet, good for financial reports and simple data entry. But it is a very poor calculator. I would not recommend even attempting to use it for ODE solutions. Even when you can, it is difficult and very opaque.
__________________
� � � � Tom Gutman

[RichardJ]

On 9/14/2009 4:36:08 PM, ArtZ wrote:

>In my world, I need to use XL
>when I can in order to share
>more easily with colleagues -
>most use XL readily and
>Mathcad hardly at all.

Then they should be beaten harder 😉

Richard

[TheodoreM.Bones]

Yes, extrapolation from a spline is not valid. It is easy to extract a portion of the lower data, curve fit that, and extrapolate somewhat from it.

Elsewhere the spline permits interpolation..

No use hand wringing.