bspline - B-Spline Interpolation and Regression
Build and use B-splines for interpolation and regression.
In case of regression, equality constraints as well as
monotonicity and/or positivity of B-spline weights can be
imposed. Moreover, knot positions can be on regular grid or be
part of optimized parameters too (in addition to the spline
weights). For this end, 'bspline' is able to calculate Jacobian
of basis vectors as function of knot positions. User is
provided with functions calculating spline values at arbitrary
points. These functions can be differentiated and integrated to
obtain B-splines calculating derivatives/integrals at any
point. B-splines of this package can simultaneously operate on
a series of curves sharing the same set of knots. 'bspline' is
written with concern about computing performance that's why the
basis and Jacobian calculation is implemented in C++. The rest
is implemented in R but without notable impact on computing
speed.
