Regression Calculator: Linear, Polynomial, Exponential, Power and Log Fits

Fits a model to your x, y data by least squares and gives the equation, R squared, adjusted R squared, residuals and a plot of the curve over the points.

Use the tool

Why this tool exists

A fitted curve turns a table of numbers into a rule you can use. Paste your x and y values, pick linear, polynomial (degree 2 to 6), exponential, power or logarithmic, and read the equation and how well it matches.

Linear and polynomial fits use a numerically stable QR least-squares solution on centered data. Exponential and power fits are done on the logarithms, the usual spreadsheet method, so their R squared is computed on the original y values and can be lower than the log-scale fit suggests. A high R squared does not prove the model is right, so always look at the plot and the residual column.

How to use it

  1. Paste one x y pair per line.
  2. Choose the model, and the degree for polynomial fits.
  3. Read the equation, R squared and the residual table.
  4. Look at the plot to see if the curve really follows the points.

Frequently asked questions

What does R squared mean?

The share of the variation in y that the model explains, from 0 to 1 for a good fit.

Why does exponential need y above zero?

The fit takes the logarithm of y, which is undefined for zero or negative values.

Can I predict new values?

Use the equation shown. Extrapolating far beyond your data is risky.

How many points do I need?

At least 3, and more than the number of model parameters.