exponential regression calculator

What this exponential regression calculator does

This calculator finds the best-fit exponential curve for your dataset using least squares regression. It models relationships of the form:

y = a · b^x   (equivalently, y = a · e^kx)

Exponential regression is useful when values change by a roughly constant percentage over time rather than a constant amount. Common examples include compound growth, population trends, biological processes, radioactive decay, and learning curves.

How to use the calculator

Step-by-step

• Enter each data point on its own line as x,y.
• Click Calculate Exponential Regression.
• Read the estimated coefficients, equation, R² values, and RMSE.
• Optionally provide a future x value to generate a prediction.

The tool also shows a small table of observed vs predicted values so you can quickly evaluate fit quality.

Understanding the output

Equation coefficients

• a: the estimated starting scale (value when x = 0 in the model).
• b: multiplicative change per 1-unit increase in x.
• k: continuous growth/decay rate where b = e^k.

Growth vs decay

• If b > 1 (or k > 0): exponential growth.
• If 0 < b < 1 (or k < 0): exponential decay.

Goodness of fit metrics

• R² (log scale): fit quality in transformed linear space ln(y).
• R² (original scale): fit quality relative to raw y values.
• RMSE: typical prediction error magnitude on y scale.

When exponential regression is appropriate

Use exponential curve fitting when your data exhibits percentage-based change. It is often better than linear regression for:

• Investment and compound interest modeling
• Population growth and epidemiology
• Battery discharge and cooling curves
• Marketing reach and adoption dynamics
• Depreciation and decay processes

Important assumptions and limitations

• All y values must be strictly positive for logarithmic transformation.
• The method assumes an exponential relationship is reasonable for the data.
• Outliers can strongly affect parameter estimates.
• Extrapolation (predicting far beyond your x-range) can be risky.

Example interpretation

Suppose you fit data and get y = 120 · 1.5^x. That means each 1-unit increase in x multiplies y by 1.5 (a 50% increase). If x is time in years, your model suggests 50% annual growth.

For decay, you might see y = 500 · 0.82^x, indicating an 18% drop per x unit.

Tips for better regression results

• Use at least 5–10 data points when possible.
• Keep units consistent for x and y.
• Plot your data to visually inspect whether an exponential pattern is plausible.
• Compare with linear or polynomial models if fit metrics are weak.