Post History
I'm spitballing a story where the main character goes through a monstrous transformation. Part of the conceit / my enjoyment is that she's an oncology researcher and will be tracking the transform...
Question
mathematics
#2: Post edited
by
trichoplax
·
2023-10-01T02:31:53Z (about 1 year ago)
Add spaces around slash where one of the alternatives is more than a single word, plus typos
Modelling a monstrous transformation - generalized logistic function fitting.
- I'm spitballing a story where the main character goes through a monstrous transformation.
Part of the conceit/my enjoyment is that she's an oncology researcher and will be tracking the transformation over its 18 month duration and trying to understand the chemical and biological processes occurring.- I've semi-arbitrarily decided her starting weight/height (90lbs and 5'6") as well as her end weight/height (1700lbs and ~9’0“). The other aspects of her transformation (teeth, snout, arm length etc) should be irrelevant.
- I wanted have the changes follow a general logistic curve, which is a more general version of the Gompertz curve (used in [modelling tumor growth](https://en.wikipedia.org/wiki/Gompertz_function#Growth_of_tumors)).
I've created a spreadsheet where I'd already modelled the transformation with a [gaussian function](https://en.wikipedia.org/wiki/Gaussian_function), and figured out (mostly my trial and error) *a*, *b* and *c* values and just added the results at each timestep to make it cumulative... But I realised recently that I can't easily skew that plot.Part of the reason I wanted it skewed, is that she'll discover a complex negative feedback loop that stops the biological processes from running away and never stoping.More concretely her height shoots up in the first 6 months, then slows down. Then her weight that's only slowly been rising also increases more rapidly before stopping abruptly. That should skew the data points she plots.- I was aiming for her to gain ⅔ of her final height in the first ⅓ of the 18 months, and ⅔ of her final weight in the last ⅓ of her 18 month transformation.
- I can almost fit her height this way, but I haven't managed to find the right parameters for her weight. I did so by looking at the starting and final weights, and then using trial and error on the generalised logistic function that's a solution to the [Richards's differential equation](https://en.wikipedia.org/wiki/Generalised_logistic_function#Generalised_logistic_differential_equation).
- Given my requirements and conditions (start and end points, as well as ⅓ and ⅔ way through points), is there a more reliable way to find the parameters for the generalised logistic function I need to plot, that isn't just trial and error?
- I'm spitballing a story where the main character goes through a monstrous transformation.
- Part of the conceit / my enjoyment is that she's an oncology researcher and will be tracking the transformation over its 18 month duration and trying to understand the chemical and biological processes occurring.
- I've semi-arbitrarily decided her starting weight/height (90lbs and 5'6") as well as her end weight/height (1700lbs and ~9’0“). The other aspects of her transformation (teeth, snout, arm length etc) should be irrelevant.
- I wanted have the changes follow a general logistic curve, which is a more general version of the Gompertz curve (used in [modelling tumor growth](https://en.wikipedia.org/wiki/Gompertz_function#Growth_of_tumors)).
- I've created a spreadsheet where I'd already modelled the transformation with a [gaussian function](https://en.wikipedia.org/wiki/Gaussian_function), and figured out (mostly by trial and error) *a*, *b* and *c* values and just added the results at each timestep to make it cumulative... But I realised recently that I can't easily skew that plot.
- Part of the reason I wanted it skewed, is that she'll discover a complex negative feedback loop that stops the biological processes from running away and never stopping.
- More concretely her height shoots up in the first 6 months, then slows down. Then her weight, that's only slowly been rising, also increases more rapidly before stopping abruptly. That should skew the data points she plots.
- I was aiming for her to gain ⅔ of her final height in the first ⅓ of the 18 months, and ⅔ of her final weight in the last ⅓ of her 18 month transformation.
- I can almost fit her height this way, but I haven't managed to find the right parameters for her weight. I did so by looking at the starting and final weights, and then using trial and error on the generalised logistic function that's a solution to the [Richards's differential equation](https://en.wikipedia.org/wiki/Generalised_logistic_function#Generalised_logistic_differential_equation).
- Given my requirements and conditions (start and end points, as well as ⅓ and ⅔ way through points), is there a more reliable way to find the parameters for the generalised logistic function I need to plot, that isn't just trial and error?
#1: Initial revision
by
pureferret
·
2023-08-07T20:39:48Z (over 1 year ago)
Modelling a monstrous transformation - generalized logistic function fitting.
I'm spitballing a story where the main character goes through a monstrous transformation. Part of the conceit/my enjoyment is that she's an oncology researcher and will be tracking the transformation over its 18 month duration and trying to understand the chemical and biological processes occurring. I've semi-arbitrarily decided her starting weight/height (90lbs and 5'6") as well as her end weight/height (1700lbs and ~9’0“). The other aspects of her transformation (teeth, snout, arm length etc) should be irrelevant. I wanted have the changes follow a general logistic curve, which is a more general version of the Gompertz curve (used in [modelling tumor growth](https://en.wikipedia.org/wiki/Gompertz_function#Growth_of_tumors)). I've created a spreadsheet where I'd already modelled the transformation with a [gaussian function](https://en.wikipedia.org/wiki/Gaussian_function), and figured out (mostly my trial and error) *a*, *b* and *c* values and just added the results at each timestep to make it cumulative... But I realised recently that I can't easily skew that plot. Part of the reason I wanted it skewed, is that she'll discover a complex negative feedback loop that stops the biological processes from running away and never stoping. More concretely her height shoots up in the first 6 months, then slows down. Then her weight that's only slowly been rising also increases more rapidly before stopping abruptly. That should skew the data points she plots. I was aiming for her to gain ⅔ of her final height in the first ⅓ of the 18 months, and ⅔ of her final weight in the last ⅓ of her 18 month transformation. I can almost fit her height this way, but I haven't managed to find the right parameters for her weight. I did so by looking at the starting and final weights, and then using trial and error on the generalised logistic function that's a solution to the [Richards's differential equation](https://en.wikipedia.org/wiki/Generalised_logistic_function#Generalised_logistic_differential_equation). Given my requirements and conditions (start and end points, as well as ⅓ and ⅔ way through points), is there a more reliable way to find the parameters for the generalised logistic function I need to plot, that isn't just trial and error?