I am completely new to ALGLIB and I was trying to perform the following task, which is to calculate K using a script with

Arthur Albano 2021.10.07 17:28 #41

Hi!

I am completely new to ALGLIB. So, I was trying to perform the following task, which is to calculate k² using a script. I can easly do that in Excel, but have no idea how to use a non-linear eqution model with one boxed constraint.

void OnStart()
  {
   const double x[22] = {28.98, 30.48, 30.98, 31.48, 31.98, 32.48, 32.98, 33.48, 33.98, 34.48, 34.98, 35.48, 35.98, 36.48, 36.98, 37.48, 37.98, 38.48, 38.98, 39.98, 40.98, 41.98};
   const double y[22] = {8.26, 7.26, 5.71, 6.15, 5.90, 5.15, 5.03, 4.30, 4.15, 3.30, 2.90, 2.71, 1.96, 1.57, 1.27, 0.90, 0.70, 0.47, 0.33, 0.20, 0.09, 0.06};
   const double w[22] = {2, 8, 2, 2, 2, 1, 11, 12, 24, 12, 6, 12, 21, 7, 74, 51, 159, 385, 171, 109, 18, 18}; //weights
   const double x0 = 37.74; //central value theorem

//model equation
//y = (1/2)*((x^2-2x0x+x0^2+4k^2)^(1/2) + x+x0) - x
//k^2 was calculated as 0.582018800686398, after using Solver in Excel where total error was a minimized sum of w(y-y0)^2
//box constraint k^2>0;
  };

Thanks in advance for reading!

[WARNING CLOSED!] Any newbie [Archive!] Any rookie question, How to use ALGLIB

Arthur Albano 2021.10.09 22:36 #42

Arthur Albano #:

Hi!

I am completely new to ALGLIB. So, I was trying to perform the following task, which is to calculate k² using a script. I can easly do that in Excel, but have no idea how to use a non-linear eqution model with one boxed constraint.

Thanks in advance for reading!

So far, looks ugly...

//+------------------------------------------------------------------+
//|                                                AlgLibExample.mq5 |
//|                                                    Arthur Albano |
//|                       https://www.mql5.com/en/users/arthuralbano |
//+------------------------------------------------------------------+
#property copyright "Arthur Albano"
#property link      "https://www.mql5.com/en/users/arthuralbano"
#property version   "1.00"
//+------------------------------------------------------------------+
//| Script program start function                                    |
//+------------------------------------------------------------------+

#include <Math\Alglib\alglib.mqh>
#include <Math\Alglib\interpolation.mqh>
#include <Arrays\ArrayDouble.mqh>

//+------------------------------------------------------------------+
//|                                                                  |
//+------------------------------------------------------------------+
void OnStart()
  {
   const double x0 = 37.74; //central value theorem

//model equation
//f(x,k²,x0) = (1/2)*((x^2-2x0x+x0^2+4k^2)^(1/2) + x+x0) - x
//k^2 was calculated as 0.582018800686398, after using Solver in Excel
//box contraint k^2>0;

// https://www.alglib.net/translator/man/manual.cpp.html#example_lsfit_d_nlfb

   const double x[22] = {28.98, 30.48, 30.98, 31.48, 31.98, 32.48, 32.98, 33.48, 33.98, 34.48, 34.98, 35.48, 35.98, 36.48, 36.98, 37.48, 37.98, 38.48, 38.98, 39.98, 40.98, 41.98};
   double w[22] = {2, 8, 2, 2, 2, 1, 11, 12, 24, 12, 6, 12, 21, 7, 74, 51, 159, 385, 171, 109, 18, 18}; //weights
   double y[22] = {8.26, 7.26, 5.71, 6.15, 5.90, 5.15, 5.03, 4.30, 4.15, 3.30, 2.90, 2.71, 1.96, 1.57, 1.27, 0.90, 0.70, 0.47, 0.33, 0.20, 0.09, 0.06};
   
   const int n = ArraySize(x);

   double c[1] = {x0 / 100.};
   CLSFitState state;

   CMatrixDouble xx(n, 1);
   for(int i = 0; i < n; i++)
     {
      xx[i].Set(0,x[i]);
     };

   CLSFit::LSFitCreateWFGH(xx, y, w, c, n,1,1,state);

   const double epsf = 0.000001;
   const double epsx = 0.000001;
   const int maxits = 0;
   CLSFit::LSFitSetCond(state,epsf,epsx,maxits);
   
   double bndl[1] = {0.};
   double bndu[1] = {x0 / 100.};
   CLSFit::LSFitSetBC(state,bndl,bndu);
   
   CNDimensional_PFunc *Pfunc = new CNDimensional_PFunc;

   //CAlglib::LSFitFit(state,Pfunc,,,rep,state);
   
   int info;
   CLSFitReport rep;
   CLSFit::LSFitResults(state,info,c,rep);
  };
//+------------------------------------------------------------------+
//|                                                                  |
//+------------------------------------------------------------------+
double func(const double _x, const double _k, const double _x0)
  {
   return((1. / 2.) * (pow(pow(_x, 2.) - 2 * _x0 * _x + pow(_x0, 2.) + 4 * pow(_k, 2.), 1. / 2.) + _x + _x0) - _x);
  };
//+------------------------------------------------------------------+

Arthur Albano 2022.01.11 02:23 #43

Done.

Basically copy & paste from AlgLib TestInterface.mqh.

//+------------------------------------------------------------------+
//|                                                   AlgLibTest.mq5 |
//|                                  Copyright 2022, MetaQuotes Ltd. |
//|                                             https://www.mql5.com |
//+------------------------------------------------------------------+
#property copyright "Copyright 2022, MetaQuotes Ltd."
#property link      "https://www.mql5.com"
#property version   "1.00"
//+------------------------------------------------------------------+
//| Script program start function                                    |
//+------------------------------------------------------------------+

#include <Math\Alglib\alglib.mqh>

//+------------------------------------------------------------------+
//|                                                                  |
//+------------------------------------------------------------------+
void OnStart()
  {
   Print(TEST_LSFit_D_NLF());
  };

//+------------------------------------------------------------------+
//| Derived class from CNDimensional_PFunc                           |
//+------------------------------------------------------------------+
class CNDimensional_CX_1_Func : public CNDimensional_PFunc
  {
public:
   //--- constructor, destructor
                     CNDimensional_CX_1_Func(void);
                    ~CNDimensional_CX_1_Func(void);
   //--- method
   virtual void      PFunc(double &c[], double &x[], double &func, CObject &obj);
  };
//+------------------------------------------------------------------+
//| Constructor without parameters                                   |
//+------------------------------------------------------------------+
CNDimensional_CX_1_Func::CNDimensional_CX_1_Func(void)
  {

  }
//+------------------------------------------------------------------+
//| Destructor                                                       |
//+------------------------------------------------------------------+
CNDimensional_CX_1_Func::~CNDimensional_CX_1_Func(void)
  {

  }
//+------------------------------------------------------------------+
//| This callback calculates f(c,x), where x is a                    |
//| position on X-axis and c is adjustable parameter                 |
//+------------------------------------------------------------------+
void CNDimensional_CX_1_Func::PFunc(double &c[], double &x[], double &func, CObject &obj)
  {
   double x0 = 37.74;
   func = (1.0/2.0)*(sqrt(x[0]*x[0] - 2.0*x[0]*x0 + x0*x0 + 4*c[0]) + x[0] + x0) - x[0];
  }

//+------------------------------------------------------------------+
//| Nonlinear fitting using function value only                      |
//+------------------------------------------------------------------+
double TEST_LSFit_D_NLF()
  {
//--- create variables
   CMatrixDouble           x;
   double                  y[];
   double                  c[];
   double                  w[];
   CObject                 obj;
   CNDimensional_CX_1_Func fcx1func;
   CNDimensional_Rep       frep;
   x.Resize(22,1);
//--- initialization
   x[0].Set(0,28.98);
   x[1].Set(0,30.48);
   x[2].Set(0,30.98);
   x[3].Set(0,31.48);
   x[4].Set(0,31.98);
   x[5].Set(0,32.48);
   x[6].Set(0,32.98);
   x[7].Set(0,33.48);
   x[8].Set(0,33.98);
   x[9].Set(0,34.48);
   x[10].Set(0,34.98);
   x[11].Set(0,35.48);
   x[12].Set(0,35.98);
   x[13].Set(0,36.48);
   x[14].Set(0,36.98);
   x[15].Set(0,37.48);
   x[16].Set(0,37.98);
   x[17].Set(0,38.48);
   x[18].Set(0,38.98);
   x[19].Set(0,39.98);
   x[20].Set(0,40.98);
   x[21].Set(0,41.98);
//--- allocation
   ArrayResize(y,22);
//--- initialization
   y[0]=8.26;
   y[1]=7.26;
   y[2]=5.71;
   y[3]=6.15;
   y[4]=5.90;
   y[5]=5.15;
   y[6]=5.03;
   y[7]=4.30;
   y[8]=4.15;
   y[9]=3.30;
   y[10]=2.90;
   y[11]=2.71;
   y[12]=1.96;
   y[13]=1.57;
   y[14]=1.27;
   y[15]=0.90;
   y[16]=0.70;
   y[17]=0.47;
   y[18]=0.33;
   y[19]=0.20;
   y[20]=0.09;
   y[21]=0.06;
//--- allocation
   ArrayResize(c,1);
//--- initialization
   c[0] =  0.5;
//--- create variables
   int maxits=0;
   int info;
   CLSFitStateShell state;
   double diffstep=0.0001;
//--- Fitting withweights
   CAlglib::LSFitCreateF(x,y,c,diffstep,state);
//--- function call
   double epsf = 0.0000001;
   double epsx = 0.0000001;
   CAlglib::LSFitSetCond(state,epsf,epsx,maxits);
//--- function call
   CAlglib::LSFitFit(state,fcx1func,frep,0,obj);
//--- function call
   CLSFitReportShell rep;
   CAlglib::LSFitResults(state,info,c,rep);
//--- Fitting with weights
//--- (you can change weights and see how it changes result)
   ArrayResize(w,22);
//--- initialization
   w[0]=2;
   w[1]=8;
   w[2]=2;
   w[3]=2;
   w[4]=2;
   w[5]=1;
   w[6]=11;
   w[7]=12;
   w[8]=24;
   w[9]=12;
   w[10]=6;
   w[11]=12;
   w[12]=21;
   w[13]=7;
   w[14]=74;
   w[15]=51;
   w[16]=159;
   w[17]=385;
   w[18]=171;
   w[19]=109;
   w[20]=18;
   w[21]=18;
//--- Clear weights
//   ArrayFill(w,0,ArraySize(w),1);
//--- function call
   CAlglib::LSFitCreateWF(x,y,w,c,diffstep,state);
//--- function call
   CAlglib::LSFitSetCond(state,epsf,epsx,maxits);
//--- function call
   CAlglib::LSFitFit(state,fcx1func,frep,0,obj);
//--- function call
   CAlglib::LSFitResults(state,info,c,rep);
   return(c[0]);
  }
//+------------------------------------------------------------------+

Yevgeniy Koshtenko 2022.06.03 23:00 #44

Please reset a fresh version, it doesn't compile!(

Renat Fatkhullin 2022.06.05 10:25 #45

Koshtenko #:
Please reset the latest version, it doesn't compile!(

Use the standard version of Alglib, which has been supplied with the terminal for a long time. The class of complex numbers is adapted in it:

Daying Cao 2022.10.14 16:51 #46

qingyouwei #:
How can i multiply two matrix use the lib alglib ?

How can i multiply two matrix use the lib alglib?

Can you compile ALGLIB and pass it?

My compiled the following error: 'complex' - unexpected token complex.mqh 37 8

'complex' - unexpected token complex.mqh 37 8

'{' - name expected complex.mqh 38 3

'complex' - unexpected token, probably type is missing? complex.mqh 79 1

'im' - unexpected token, probably type is missing? complex.mqh 79 31

'im' - class type expected complex.mqh 79 31

'}' - semicolon expected complex.mqh 82 3

'}' - expressions are not allowed on a global scope complex.mqh 82 3

'complex' - unexpected token, probably type is missing? complex.mqh 86 1

'im' - unexpected token, probably type is missing? complex.mqh 86 41

'im' - variable already defined complex.mqh 86 41

see previous declaration of variable 'im' complex.mqh 79 31

'im' - class type expected complex.mqh 86

how to solve unexpected ... Unexpected token, probably MQL5 : Header (mqh)

Ivan Shishikin 2023.04.03 15:32 #47

When are we going to add spectral analysis functions
SSA ?
ALGLIB has them, but MT still does not.
https://www.alglib.net/translator/man/manual.cpp.html#example_ssa_d_basic

Stanislav Korotky 2023.04.03 20:59 #48

Ivan Shishikin #:

When are we going to add spectral analysis functions
SSA ?
ALGLIB has them, but MT still does not.
https://www.alglib.net/translator/man/manual.cpp.html#example_ssa_d_basic

Maybe until the standard library will not be added (and it may not be soon), the existing sources(1, 2) will be fine?

Ivan Shishikin 2023.04.04 10:03 #49

unfortunately there is no code for predicates.
there is only the model itself. but that is not enough.

Denis Kirichenko 2023.04.04 10:22 #50

Ivan Shishikin #:

When are we going to add spectral analysis functions
SSA ?
ALGLIB has them, but MT still does not.
https://www.alglib.net/translator/man/manual.cpp.html#example_ssa_d_basic

Imho, there is no guarantee that it will appear at all ... Maybe, instead of waiting, it is better to use them directly in C-code?

Libraries: ALGLIB - Numerical Analysis Library - page 5