Exponential Regression Calculator

Exponential regression is a type of curve-fitting technique used to find the exponential function that best describes a set of data points. It models relationships where the rate of change is proportional to the current value, such as population growth, radioactive decay, or compound interest. The resulting model takes the form y = a·bˣ, where a and b are constants determined from the data.

Exponential regression is solved by linearizing the model. Taking the natural logarithm of both sides gives ln(y) = ln(a) + x·ln(b), which is a standard linear equation. Ordinary least squares (OLS) regression is then applied to the transformed data (x, ln(y)) to find the slope and intercept, from which a = e^(intercept) and b = e^(slope) are recovered. This calculator performs all these steps automatically.

R-squared (R²) measures how well the exponential model explains the variation in your data. A value of 1.0 means a perfect fit, while 0 means the model explains none of the variability. In exponential regression, R² is computed on the log-transformed values, so it reflects goodness of fit in the linearized space. Values above 0.9 generally indicate a strong exponential relationship.

Linear regression fits a straight line (y = a + bx) and assumes the response variable changes by a constant amount per unit increase in x. Exponential regression fits a curve (y = a·bˣ) and assumes the response variable changes by a constant percentage per unit increase in x. Exponential regression is appropriate when your data grows or decays at an accelerating rate, rather than at a steady additive pace.

You need a minimum of 3 data points to fit an exponential regression model, since the calculation involves estimating two parameters (a and b). However, more data points produce a more reliable and stable model. As a practical guideline, 10 or more data points are recommended for meaningful statistical conclusions.

This calculator uses the exponential model y = a·bˣ, which can also be written as y = a·e^(cx) where c = ln(b). The parameter a is the y-intercept (value of y when x = 0), and b (or e^c) controls the growth rate. If b > 1, the model represents exponential growth; if 0 < b < 1, it represents exponential decay.

Exponential regression requires taking the natural logarithm of the y values during the linearization step. Since the logarithm is only defined for positive numbers, all y values must be strictly greater than zero. If your data contains zero or negative y values, an exponential model is not directly applicable and you should consider a different regression type.

In the model y = a·bˣ, the coefficient a represents the starting value — the predicted y when x equals zero. The coefficient b is the growth (or decay) factor per unit increase in x. For example, if b = 1.5, the y value multiplies by 1.5 for each one-unit increase in x, representing 50% growth per step. If b = 0.8, the y value shrinks to 80% per step, representing 20% decay.