This function is used to calculate the estimates of \(\mu(t), \beta_i(s,t), \gamma_{ij}(u,v,t)\) for function-on-function interaction model (see the description in cv.ff.interaction) based on the output object of cv.ff.interaction, or step.ff.interaction.
getcoef.ff.interaction(fit.obj, t.x.coef=NULL, t.y.coef=NULL)
the output object of cv.ff.interaction, or step.ff.interaction
a list of length \(p\) of vectors providing the observation time
points of predictors on which coefficient functions will be
evaluated. If t.x.coef=NULL (default), t.x in
cv.ff.interaction or step.ff.interaction
will be used.
a vector of observation time points of response
function on which the coefficient functions will be evaluated. If
t.y.coef=NULL (default), t.y in
cv.ff.interaction or step.ff.interaction
will be used.
a list providing the given or selected main effects and interactions, together with the corresponding estimated coefficient functions.
the vector of estimated \(\mu(t)\) evaluated at the vector t.y.coef of the observation points for the response function \(y(t)\).
the index vector of the input main_effects for cv.ff.interaction or the selected main effects by step.ff.interaction.
a list of matrices of the estimated values of the coefficient functions of main effect specified by main_effects. Each matrix gives the estimated values of \(\beta_i(s,t)\) at the two-dimensional grid created by the observation point vectors t.x.coef[[i]] and t.y.coef, where \(i\) is an index in main_effects.
a matrix of two columns showing the input interactions for cv.ff.interaction or the selected interactions by step.ff.interaction. Each row shows the indices of the pair of functional variables in an interaction or quadratic effect.
a list of three-dimensional arrays of estimated values of the coefficient functions of interaction or quadratic effects specified by inter_effects. Each array gives the estimated values of \(\gamma_{ij}(u,v,t)\) at the three-dimensional grid created by the observation point vectors t.x.coef[[i]], t.x.coef[[j]] and t.y.coef, where the pair \(i, j\) is in inter_effects.
Ruiyan Luo and Xin Qi (2018) Interaction model and model selection for function-on-function regression, Journal of Computational and Graphical Statistics. https://doi.org/10.1080/10618600.2018.1514310
# NOT RUN {
#See the examples in cv.ff.interaction() and step.ff.interaction().
# }
Run the code above in your browser using DataLab