Calculator Inputs
Example Data Table
| x | y | Meaning |
|---|---|---|
| 1 | 12 | First observed value |
| 2 | 18 | Early growth value |
| 3 | 27 | Increasing trend value |
| 4 | 40 | Middle sample value |
| 5 | 61 | Higher measured value |
| 6 | 90 | Late growth value |
| 7 | 136 | Final observed value |
Formula Used
The calculator fits this exponential model:
y = ae^(kx)
It also shows the equivalent base form:
y = ab^x
First, the data is transformed with the natural logarithm:
ln(y) = ln(a) + kx
The slope and intercept are calculated with linear least squares:
k = [nΣ(x ln(y)) - ΣxΣln(y)] / [nΣx² - (Σx)²]
ln(a) = [Σln(y) - kΣx] / n
Then:
a = e^(ln(a)) and b = e^k
Predicted values are calculated with:
ŷ = ae^(kx)
Residuals are calculated with:
Residual = y - ŷ
How to Use This Calculator
- Enter positive data pairs in the large input box.
- Place one x,y pair on each line.
- Enter the x value where you want a prediction.
- Select the number of decimal places.
- Keep sorting enabled for a cleaner graph.
- Press the calculate button.
- Review the equation, fit scores, chart, and residual table.
- Use the CSV or PDF buttons to export your result.
Exponential Regression Guide
What It Means
Exponential regression is useful when change speeds up or slows down by a constant ratio. It is common in growth studies, decay studies, finance, biology, and web traffic analysis. The model turns curved data into a straight line by taking the natural log of every positive y value.
Main Equation
This calculator fits the equation y = a e^(kx). The value a is the starting scale. The value k is the continuous growth rate. It also shows the equivalent form y = a b^x, where b equals e^k. A b value above one shows growth. A b value below one shows decay.
Data Quality
Good data matters. Every y value must be greater than zero, because logarithms of zero or negative values are not defined. Points should come from one process. Mixed processes can create weak fits and misleading predictions. Use the residual table to find unusual points. Large residuals may show errors, outliers, or a model that is too simple.
Fit Scores
The graph compares measured values with the fitted curve. A close curve suggests a useful equation. R squared shows how much variation is explained by the model. RMSE shows the typical prediction error in the original y units. MAPE gives an average percentage error, which is easy to understand for many business cases.
Predictions
Prediction is simple after fitting the curve. Enter an x value, and the tool estimates y from the regression equation. Use predictions inside the range of your data when possible. Far outside values are extrapolations. They can become risky, especially when growth changes over time.
Advanced Review
Advanced users can compare log fit and original fit. The log fit controls the estimated parameters. The original fit checks real scale accuracy. Review both values before making decisions. When errors grow with larger y values, the logarithmic approach is often helpful. When errors stay equal across all y values, another model may be better for your data set.
Exports
CSV and PDF exports help you keep records. The CSV file works well in spreadsheets. The PDF file is useful for reports and class work. Always include the original data, equation, fit scores, and residuals when sharing results. This makes your regression clear, repeatable, and easier to review.
FAQs
What is an exponential regression equation?
It is an equation that fits data following a curved growth or decay pattern. The common form is y = ae^(kx), where a controls scale and k controls growth or decay rate.
Why must y values be positive?
The method uses ln(y). Natural logarithms are only defined for positive values in this real number calculation. Zero or negative y values cannot be transformed correctly.
What does the k value mean?
The k value is the continuous growth rate. A positive k means exponential growth. A negative k means exponential decay. A k near zero shows little exponential change.
What does the b value mean?
The b value is the multiplier for each one unit increase in x. If b is 1.2, the fitted value grows by about 20 percent per x unit.
What is R squared?
R squared measures how well the curve explains variation in the data. Values closer to one usually mean a stronger fit. It should still be reviewed with residuals.
What is RMSE?
RMSE is the root mean squared error. It shows the typical size of prediction error in the same unit as y. Lower RMSE usually means better accuracy.
Can I use this for decay data?
Yes. Decay data works when all y values are positive. The calculator will return a negative k value and a b value below one for decay trends.
Are predictions outside the data range safe?
They are extrapolations, so they are less reliable. Exponential equations can grow or shrink quickly. Use outside-range predictions with caution and domain knowledge.