Bounded optimal knots for regression splines

Using a B-spline representation for splines with knots seen as free variables, the approximation to data by splines improves greatly. The main limitations are the presence of too many local optima in the univariate regression context, and it becomes even worse in multivariate additive modeling. When the number of knots is a priori

• Type: Journal
• Year: 2004
• Where: CSDA
• Authors: Nicolas Molinari, Jean-François Durand, Robert Sabatier
• Keywords: CSDA 2004, Least-squares Spline Approximations, Multivariate Additive Modeling, Univariate Regression Context