122 JParser<> zap(
"Program to test JGandalf algorithm.");
132 catch(
const exception& error) {
133 FATAL(error.what() << endl);
136 using namespace JFIT;
138 ASSERT(numberOfEvents > 0);
142 TF1 fs(
"fs",
"exp(-0.5 * (x-[0])*(x-[0]) / ([1]*[1]))");
145 fs.FixParameter(0,
gauss[0]);
146 fs.FixParameter(1,
gauss[1]);
149 const Double_t xmin = -5.0;
150 const Double_t xmax = +5.0;
157 TH1D
H[] = { TH1D(
"ha",
"", 101, -0.1, +0.1),
158 TH1D(
"hb",
"", 101, -0.1, +0.1),
159 TH1D(
"hc",
"", 101, -100.0, +100.0),
160 TH1D(
"hd",
"", 101, -100.0, +100.0) };
164 for (
int i = 0; i != numberOfEvents; ++i) {
166 STATUS(
"event: " << setw(10) << i <<
'\r');
DEBUG(endl);
168 TH1D h0(
"h0", NULL, nx, xmin, xmax);
172 h0.FillRandom(
"fs", (Int_t)
gauss[2]);
173 h0.FillRandom(
"fb", (Int_t) gauss[3]);
177 for (Int_t i = 1; i <= h0.GetNbinsX(); ++i) {
178 data.push_back(JElement_t(h0.GetBinCenter (i),
179 h0.GetBinContent(i)));
186 for (
size_t i = 0; i != 4; ++i) {
192 h0.GetEntries() * (xmax - xmin) / nx - h0.GetMinimum(),
199 const double chi2 = fit(
g1, data.begin(), data.end());
204 DEBUG(
"Chi2 " << chi2 << endl);
206 const double Y[] = { fit.
value[0] - gauss[0],
207 fit.
value[1] - gauss[1],
208 fit.
value[2] * nx / (xmax - xmin) - gauss[2],
209 fit.
value[3] * nx - gauss[3] };
211 for (
int i = 0; i !=
sizeof(Q)/
sizeof(Q[0]); ++i) {
217 for (
int i = 0; i !=
sizeof(Q)/
sizeof(Q[0]); ++i) {
222 timer.print(cout,
true,
micro_t);
229 for (
int i = 0; i !=
sizeof(
H)/
sizeof(
H[0]); ++i) {
237 for (
int i = 0; i !=
sizeof(Q)/
sizeof(Q[0]); ++i) {
241 ASSERT(Q[0].getSTDev() < precision[0]);
242 ASSERT(Q[1].getSTDev() < precision[1]);
243 ASSERT(Q[2].getSTDev() < precision[2]);
244 ASSERT(Q[3].getSTDev() < precision[3]);
Utility class to parse command line options.
std::vector< parameter_type > parameters
double getMean(vector< double > &v)
get mean of vector content
#define ASSERT(A,...)
Assert macro.
#define make_field(A,...)
macro to convert parameter to JParserTemplateElement object
Auxiliary class for CPU timing and usage.
Fit method based on the Levenberg-Marquardt method.
std::ostream & longprint(std::ostream &out)
Set long printing.
double gauss(const double x, const double sigma)
Gauss function (normalised to 1 at x = 0).
std::ostream & shortprint(std::ostream &out)
Set short printing.
#define DEBUG(A)
Message macros.
Double_t g1(const Double_t x)
Function.