How Exponential Regression Works
Exponential regression finds the best-fit curve of the form y = ab^x for your data. It is used to model exponential growth (population, investments, viral spread) or decay (radioactive material, depreciation).
| R² Value | Fit Quality | Interpretation |
|---|---|---|
| 0.95 – 1.00 | Excellent | Strong exponential relationship |
| 0.80 – 0.94 | Good | Reasonably well-fitted curve |
| 0.60 – 0.79 | Moderate | Some exponential trend present |
| Below 0.60 | Weak | Try linear or power regression |
Frequently Asked Questions
Use it when your data increases or decreases at a proportional rate per unit. Real-world examples: bacterial colony growth, compound interest portfolios, radioactive decay, viral social media posts, technology adoption curves, and COVID-19 case counts early in an outbreak.
Exponential regression requires all y values to be strictly positive because we take ln(y). If your data includes zeros or negatives, you cannot use exponential regression directly. Consider shifting data upward or using a different model.
The base b is the multiplicative growth or decay factor per unit increase in x. If b > 1, y grows exponentially. If 0 < b < 1, y decays. For example, b = 2.718 means y approximately triples per unit x.
Linear regression fits y = mx + b (additive change — constant increase per unit x). Exponential regression fits y = ab^x (multiplicative change — proportional increase per unit x). Exponential models apply when percent change is constant.
R² above 0.95 is excellent. Above 0.80 is good. Below 0.60 suggests the data does not follow an exponential pattern well — try comparing with linear or power regression.
Yes, but with caution. Exponential regression is useful for short-term forecasting of exponentially-trending data. However, true exponential growth cannot continue indefinitely — long-range forecasts should be treated as upper bounds.