Rev 6817 | Blame | Compare with Previous | Last modification | View Log | Download | RSS feed
/*Routines for fitting gnls modelsCopyright 1997-2005 Douglas M. Bates <bates@stat.wisc.edu>,Jose C. Pinheiro,Saikat DebRoyCopyright 2007-2012 The R Core TeamThis file is part of the nlme package for R and related languagesand is made available under the terms of the GNU General PublicLicense, version 2, or at your option, any later version,incorporated herein by reference.This program is distributed in the hope that it will beuseful, but WITHOUT ANY WARRANTY; without even the impliedwarranty of MERCHANTABILITY or FITNESS FOR A PARTICULARPURPOSE. See the GNU General Public License for moredetails.You should have received a copy of the GNU General Public Licensealong with this program; if not, a copy is available athttp://www.r-project.org/Licenses/*/#include "nlOptimizer.h"#include "matrix.h"#include "nlmefit.h"extern void corStruct_recalc(double *, longint *, longint *, double *);/* gnls functions and variables */typedef struct gnls_struct { /* Generalized nonlinear least squares structure */double *residuals, *gradient, *corFactor, *varWeights, minFactor,tolerance, *newtheta, *theta, *incr, *add_ons,new_objective, objective;double *result[1];longint corOpt, varOpt, npar, ncol, N, nrdof, maxIter, *corDims;#ifdef R_S_HSEXP model;#endifint conv_failure;} *gnlsPtr;static gnlsPtrgnls_init(double *ptheta, longint *dims, double *corFactor, double *varWeights,longint *corDims, double *settings, double *additional,longint corOpt, longint varOpt aMOD){longint nResult;gnlsPtr gnls = Calloc(1, struct gnls_struct);gnls->theta = ptheta;gnls->corFactor = corFactor;gnls->varWeights = varWeights;gnls->corDims = corDims;gnls->npar = dims[0];gnls->N = dims[1];gnls->nrdof = gnls->N - gnls->npar;gnls->ncol = gnls->npar + 1;gnls->maxIter = (int) settings[0];gnls->minFactor = settings[1];gnls->tolerance = settings[2];gnls->newtheta = Calloc(gnls->npar, double);gnls->incr = Calloc(gnls->npar, double);gnls->varOpt = varOpt;gnls->corOpt = corOpt;gnls->add_ons = additional;#ifdef R_S_Hgnls->model = model;gnls->result[0] = DNULLP;nResult = evaluate(ptheta, gnls->npar MOD, gnls->result SEV);gnls->result[0] = Calloc(nResult, double);#endifreturn gnls;}static voidgnlsFree( gnlsPtr gnls ){Free(gnls->newtheta);Free(gnls->incr);#ifdef R_S_HFree(gnls->result[0]);#endifFree(gnls);}static doublegnls_objective(gnlsPtr gnls){longint i, j;if(gnls->varOpt) { /* variance function correction */for(i = 0; i < gnls->N; i++) {for(j = 0; j < gnls->ncol; j++) {*(gnls->result[0] + i + j * gnls->N) *= gnls->varWeights[i];}}}if(gnls->corOpt) { /* correlation structure correction */corStruct_recalc(gnls->result[0], gnls->corDims, &gnls->ncol, gnls->corFactor);}gnls->residuals = gnls->result[0] + gnls->npar * gnls->N;gnls->gradient = gnls->result[0];return(d_sum_sqr(gnls->residuals, gnls->N));}/* static double *//* gnls_RegSS(gnlsPtr gnls) *//* { *//* longint i; *//* double regSS = 0, aux; *//* for(i = 0; i < gnls->N; i++) { *//* aux = d_dot_prod(gnls->gradient + i, gnls->N, gnls->incr, 1L, gnls->npar); *//* regSS += aux * aux; *//* } *//* return(regSS); *//* } */static doublegnls_increment(gnlsPtr gnls){double regSS = 0, *auxRes;QRptr aQR;longint i;if (!sqrt_eps) sqrt_eps = sqrt(DOUBLE_EPS);auxRes = Calloc(gnls->N, double);Memcpy(auxRes, gnls->residuals, gnls->N);aQR = QR(gnls->gradient, gnls->N, gnls->N, gnls->npar);QRsolve(aQR, gnls->residuals, gnls->N, 1L, gnls->incr, gnls->npar);QRqty(aQR, auxRes, gnls->N, 1L);for(i=0; i < gnls->npar; i++) {regSS += auxRes[i] * auxRes[i];}QRfree(aQR);Free(auxRes);return(sqrt(((double) gnls->nrdof) * regSS /((double) gnls->npar) * (gnls->new_objective - regSS)));}static longintgnls_iterate(gnlsPtr gnls aSEV){double factor, criterion;longint iteration;#ifdef R_S_HSEXP model = gnls->model;#endifS_EVALUATORMemcpy(gnls->newtheta, gnls->theta, gnls->npar);evaluate(gnls->theta, gnls->npar MOD, gnls->result SEV);gnls->new_objective = gnls->objective = gnls_objective(gnls);gnls->conv_failure = 0;for (factor = 1.0, iteration = 1; iteration <= gnls->maxIter;iteration++) { /* outer iteration loop *//* increment and convergence criterion */criterion = gnls_increment(gnls);if (gnls->conv_failure) return(iteration); /* Unable to make increment */if (criterion < gnls->tolerance) return(iteration); /* successful completion */do { /* inner loop for acceptable step size */if (factor < gnls->minFactor) {gnls->conv_failure = 1;return(iteration);}Memcpy(gnls->newtheta, gnls->theta, gnls->npar);d_axpy(gnls->newtheta, factor, gnls->incr, gnls->npar);evaluate(gnls->newtheta, gnls->npar MOD, gnls->result SEV);gnls->new_objective = gnls_objective(gnls);if (gnls->conv_failure) return(iteration); /* unable to evaluate objective */factor /= 2.0;} while (gnls->new_objective >= gnls->objective);factor *= 4.0;if (factor > 1.0)factor = 1.0;gnls->objective = gnls->new_objective;Memcpy(gnls->theta, gnls->newtheta, gnls->npar);}gnls->conv_failure = 2; /* Maximum number of iterations exceeded */return(iteration - 1);}static voidgnls_wrapup(gnlsPtr gnls aSEV){#ifdef R_S_HSEXP model = gnls->model;#endifS_EVALUATORevaluate(gnls->theta, gnls->npar MOD, gnls->result SEV);Memcpy(gnls->add_ons, gnls->result[0] + gnls->npar * gnls->N, gnls->N);gnls->objective = gnls_objective(gnls);}voidfit_gnls(double *ptheta, longint *pdims, double *pcorFactor, double*pvarWeights, longint *pcorDims, double *settings,double *additional, longint *pcorOpt, longint *pvarOpt aMOD){gnlsPtr gnls;S_EVALUATOR#ifdef R_S_HPROTECT(model);#endif /* R_S_H */if(!sqrt_eps) sqrt_eps = sqrt(DOUBLE_EPS);gnls = gnls_init(ptheta, pdims, pcorFactor, pvarWeights, pcorDims,settings, additional, *pcorOpt, *pvarOpt MOD);settings[4] = (double) gnls_iterate(gnls SEV);gnls_wrapup(gnls SEV);settings[3] = gnls->conv_failure;settings[5] = gnls->objective;gnlsFree(gnls);#ifdef R_S_HUNPROTECT(1);#endif /* R_S_H */}