Once again, you are attempting to shove a square peg into a round hole. Once the pandemic starts to break and the curve reaches an inflection point, my first indication is going to be a DECREASE in the R squared value, as the data points no longer land on the curve. It won’t cause an increase to values that are already 0.98-0.99. At that time, and ONLY at that time would it be worth the effort to derive an exponential function to describe the curve. Why on earth would I do all of that work now, when I already have years of experience telling me that the polynomial gives values within a few thousandths of the real values, well within any margin of error?
I get the feeling that you’re trying to demonstrate that you have more knowledge than an actual expert in the field. You don’t. The desire to do that is well-known as the Dunning-Kroger effect.
Can we stop already with bickering laced sexual innuendo???
Here's a tip for you. Take the problem and work backwards...(no pun intended).
Like I said, you don’t have a good model for predicting the future. If you did people wouldn’t need to go through the exercise.
You are applying a tool to data in an attempt to understand things and that’s applaudable, but fitting data has limits. You can’t interpolate a data set and expect to extrapolate from that fit. It almost never works, unless you have a good model to fit with already.
I say again, and only because you insist on making the point more evident with every one of your posts, you don’t even understand the excel tool you are using. And you clearly don’t understand least square fitting and what the R residual is.
Where do you think excel is getting the coefficients for those poly curves? Here’s a test ms PhD:
y = observables
x = variables
A = matrix of coefficients that get you from x to what you see in y
Tell us how you solve for A. That’s what excel has to do to give you your curves. Hint: it’s called least square fitting and it involves solving for the pseudo inverse of A.
Secondly, what is the error between the data set and the least square fit curve excel just solve. Hint: it’s related to R. Bigger hint: it is R. So little R means decent fit. Bigger R means fit is getting worse.
Finally, in one post you said once you see the curve break you would then derive the exponential equation. Really? Please explain how you are going to ‘derive’ that. Change the function in excel to include the exponential? That is not deriving, by the way.
I don’t try to tell you how a virus or cell multiplies. Please don’t try to tell me how you fit data to models. You are out of your lane. But don’t feel bad. Many so called experts in various science fields have grown up using tools they don’t understand in inappropriate applications and claiming they can predict behaviors that they really can’t. Climate scientists are one example where they have grossly misrepresented science and data by doing exactly that.
I encourage you, as a person who should be interested in understanding the data they are studying, to listen to experts in other fields who might be able to enlighten you to different methods or ways to look at things. I majored, both undergrad and postgrad in engineering, but I studied with some brilliant mathematicians and probably learned more from them about how to analyze problems than any other discipline I have run across. Mathematics is universal. It doesn’t lie. But statistics and assumptions are easily manipulated to get almost any answer you want. Ask a pollster.