Modelling an Exponential Curve

Click For Summary
SUMMARY

The discussion centers on modeling an exponential curve using the equation format y = a * b^t. The user, ProPM, identifies the initial value a as 554.8, derived from the first y-coordinate when t = 0. To determine the growth rate b, ProPM considers using the ratio of subsequent y-coordinates, specifically (657.5/609). A suggestion is made to utilize the natural logarithm to transform the equation into a linear form for easier calculation, and the least squares method is recommended for handling data with noise.

PREREQUISITES
  • Understanding of exponential functions and their properties
  • Familiarity with logarithmic transformations
  • Basic knowledge of linear equations
  • Introduction to least squares regression analysis
NEXT STEPS
  • Learn how to apply natural logarithms to exponential equations
  • Study the least squares method for curve fitting
  • Explore advanced techniques in regression analysis
  • Investigate tools for data visualization to plot exponential curves
USEFUL FOR

Mathematicians, data analysts, and anyone involved in modeling growth patterns or analyzing datasets that exhibit exponential behavior.

ProPM
Messages
64
Reaction score
0
Hi guys,

I need to model a curve and I am trying to come up with an exponential equation. I have several points I can work with and this is what I tried so far. (I want to know if this is the correct method or if there is a more precise way to go about this)

I am working on something of the format: a*bt

My first point in the graph will be when t = 0, hence, I presume the best pick for my value of a would be my first y coordinate point in the graph: 554.8

My doubt arises from this point on:

When I have 554.8*bt should I set up a pair or simultaneous equations or can I say that b is equal to the "rate of growth of my graph"? That is, if my y coordinate points are 609, 657.5 my rate of growth would be (657.5/609)

Since my graph is not a perfect exponential fit I was thinking of adding a constant:

a*bct

But that's something different.

Thanks,
ProPM
 
Physics news on Phys.org
I'd probably use y = aebt and use two points to determine a and b. A trick is to take the natural log of both sides so you end up with a linear equation. A more elaborate method (if the points have noise and don't fit a curve perfectly) would be a least squares approach, but that's much more involved (especially in the category of precalculus).
 

Similar threads

  • · Replies 8 ·
Replies
8
Views
3K
  • · Replies 12 ·
Replies
12
Views
3K
Replies
4
Views
3K
  • · Replies 11 ·
Replies
11
Views
4K
  • · Replies 24 ·
Replies
24
Views
3K
  • · Replies 1 ·
Replies
1
Views
2K
  • · Replies 8 ·
Replies
8
Views
2K
  • · Replies 2 ·
Replies
2
Views
7K
Replies
3
Views
2K
Replies
10
Views
2K