Function that performs the actual fitting of the expert's distribution
via quadratic programming (using the solve.QP function
from the quadprog package). This function is mainly for
internal use.
Usage
fit.SEL(N, alpha, P, A, b, gamma, dplus = 0)
Arguments
N
Matrix containing the B-spline basis functions evaluated at
the elicited quantiles.
alpha
Vector giving the levels of the elicited quantiles.
P
Penalty matrix (called Omega in the reference below).
A,b
Matrix and vector containing the specifying the constraints A'b >= 0.
gamma
Gamma parameter trading of goodness of fit and
Brier entropy/Brier divergence.
dplus
Additional offset for dvec in solve.QP.
Value
Returns solution of the quadratic programming problem.
Author(s)
Bjoern Bornkamp
References
Bornkamp, B. and Ickstadt, K. (2009). A Note on B-Splines for
Semiparametric Elicitation. The American Statistician,
63, 373–377
Goldfarb, D., and Idnani, A. (1982), Dual
and Primal-Dual Methods for Solving Strictly Convex Quadratic
Programs, in Numerical Analysis, (eds.) J. Hennart,
Springer Verlag, Berlin, pp. 226–239.
Goldfarb, D., and Idnani, A. (1983), A
Numerically Stable Dual Method for Solving Strictly Convex Quadratic
Programs”, Mathematical Programming, 27, 1–33.