testStat {tvcure} | R Documentation |
Significance test of an additive term
Description
Significance test of an additive term relying on the methodology in Wood (Biometrika 2013). It is extracted from a hidden function in the 'mgcv' package. The additive term is estimated using the product of a matrix <X> and a vector <p>.
Usage
testStat(p, X, V, rank = NULL, type = 0, res.df = -1)
Arguments
p |
Vector of spline parameters used to estimate f(x) |
X |
(Centered) B-spline basis evaluated on a fine regular grid on the support of variable <x> |
V |
Posterior variance-covariance matrix of parameter <p> |
rank |
Effective dimension of <p> |
type |
0 value by default |
res.df |
-1 indicates that the scale is fixed (cf. ordinal response) |
Value
Returns a list with following elements:
stat : Value of the test statistics
pval : P-value of the test for the null hypothesis Ho: p=0
rank : Effective dimension of <p>
[Package tvcure version 0.6.6 Index]