Linear Fitting
Linear fitting (regression analysis) gives a line or simple curve that best fits the data.
Unlike nonlinear curve fitting, linear regression does not require initial values.
Even if the data trend is not linear, a linearizable function which can be transformed into linear function (y = ax + b) may agree the trend well. Having two parameters, a linearizable function may not be very flexible. Nevertheless, if the trend is monotonous and not S-shaped, a linearizable function may be satisfactory.
XYGrapher supports the following linearizable functions:
- y = exp(ax + b)
- y = aln(x) + b
- y = axb
- y = ax0.5 + b
- y = a / x + b