Goodness-of-Fit Statistics

·     

 TableCurve 2D uses four common goodness of fit statistics. In the following descriptions, SSM is the sum of squares about the mean, SSE is the sum of squared errors (residuals), n is the total number of data values, and m is the number of coefficients in the model. DOF, the degree of freedom, is n-m.

 Coefficient of Determination (r-squared)

 rē = 1 - SSE/SSM

 Degree of Freedom Adjusted Coefficient of Determination

 DOF rē = (1 - SSE*(n-1))/(SSM*(DOF-1))

 Fit Standard Error (Root MSE)

 StdErr = sqrt(SSE/DOF)

 F-statistic

 F-stat = ((SSM-SSE)/(m-1))/(SSE/DOF)

 As a fit becomes more ideal, the rē values approach 1.0 (0 represent a complete lack of fit), the standard error decreases toward zero, and the F-statistic goes toward infinity.