fitcriteria: Fit Criteria for Free-Knot Splines

In freeknotsplines: Algorithms for Implementing Free-Knot Splines

Description

These functions compute various criteria for determining the fit of a free-knot spline. AIC.freekt computes the Akaike Information Criterion, with k determining the amount of the penalty. AICc.freekt computes the corrected Akaike Information Criterion. BIC.freekt computes the Bayesian Information Criterion, also known as Schwarz Information Criterion. adjAIC.freekt computes an adjusted Akaike Information Criterion with the penalty increased to account for the greater flexibility of free knots. adjGCV.freekt computes an adjusted GCV with the degrees of freedom increased to account for the greater flexibility of free knots.

Usage

## S3 method for class 'freekt'
AIC(object, ..., k = 2)
AICc.freekt(object)
## S3 method for class 'freekt'
BIC(object, ...)
adjAIC.freekt(object)
adjGCV.freekt(object, d = 3)

Arguments

object	An object of class "freekt" obtained by using one of the fitting algorithms.
k	The amount of the penalty. Used only for AIC.freekt.
d	The amount of the penalty. Used only for adjGCV.freekt.
...	Additional arguments to be passed to the AIC.freekt and BIC.freekt functions.

Value

Returns the value of the specified fit criterion.

Author(s)

Steven Spiriti

References

Spiriti, S., Eubank, R., Smith, P., Young, D., "Knot Selection for Least-Squares and Penalized Splines," Journal of Statistical Computation and Simulation, in press.

See Also

fit.search.numknots, which uses these fit criteria to determine the number of knots.

freeknotsplines documentation built on May 2, 2019, 8:51 a.m.