Linear and non-linear regression in gnuplot

You can fit a simple linear regression to a set of data quickly using gnuplot:

# Fit a linear polynomial to the data.
# These are the primary instructions for linear regression.

    f(x) = a*x + b                   # define a linear function

    fit f(x) 'test.dat' via a,b      # compute the regression coefficients a,b

    plot f(x), 'test.dat'            # plot the line and the data

It can also apparently do non-linear fitting, but I've not tested it:

# Fit a 2nd degree polynomial to the data.

    f(x) = a*x**2 + b*x + c          # define a quadratic function

    fit f(x) 'test.dat' via a,b,c    # compute the coefficients a,b,c

    plot f(x), 'test.dat'            # plot the polynomial and the data

This comes from here.