Overview
Exponential regression fits a curve that grows or shrinks by a steady percentage, such as compound growth, population or decay.
Paste your x-y points, one per line. Every y value must be above 0. Add an x value if you want a prediction.
Worked examples
| Inputs | Result |
|---|---|
| Your x-y points0, 3.1 · 1, 4.4 · 2, 6.3 · 3, 9.2 · 4, 12.8 · 5, 18.6 | y = 3.0908 · e^(0.35831x)R²: 0.9997 from 6 pointsGrowth per step of x: 43.09% |
How it works
The fit draws a straight line through x and ln(y) by least squares, then converts it back into a curve. R² here is measured on your original y values, so it shows how well the curve itself fits.
The formula
| a | The value of y when x is 0 |
| b | The growth rate: positive grows, negative decays |
Practical tips
- If the points fall on a straight line when y is on a log scale, an exponential fit suits them.
- Real growth rarely stays exponential for long, so be careful predicting far ahead.
- Zero or negative y values can't be used, because their logarithm doesn't exist.
Frequently asked questions
What does a negative b mean?
Decay: y shrinks by a steady percentage as x increases.
Why must y be above 0?
The method takes the logarithm of y, which only exists for positive numbers.
How is growth per step worked out?
It is e^b − 1, the percentage y grows for each increase of 1 in x.