CFIT/W

If you need to interpolate among known data, CFIT/W is your tool!

Curve fitting involves, given a set of discrete data, finding a mathematical function that provides an accurate approximation to the continuum of the data. Once an appropriate mathematical description has been arrived at, the function then allows for further points to be interpolated between the defining points.

CFIT/W can be distinguished from other curve fitting software in it's approach to the fidelity of the approximation it produces. Instead of minimizing the error of the approximation to the data at the defining points, CFIT/W allows the user to find a curve fit where the error is zero at the defining points. Available curve fits include:

• Piecewise Linear
• Least Squares
• Polynomial
• Rational Function
• Cubic Spline
• Exponential Tension Spline
• Rational Tension Spline
• Akima Spline
• Conic Spline
• Cubic B-Spline
• Monotone Spline
• Bicubic Spline
• Biquintic Spline
• Biquadratic
• Fourier Series

On many of these curve fits, the user can also specify one of three end conditions:

• Natural (zero second derivative)
• Clamped (specified first derivative)
• Bessel (first derivative from local quadratic curve fit)

CFIT/W was created to allow the user to visually inspect the quality of a curve fit. This is necessary because a type of function that fits on set of data will not necessarily do a good job of fitting another set of data. Experience has shown that a visual inspection is often the best way to judge the quality of a curve fit.