/* DENPLOT.GPR: procedure to compute univariate kernel density estimator at a vector 
                of points, and then plot results

	        calls kden.g to compute kernel density estimate at points to be plotted

                John Rust, University of Maryland, October 2006
	   
	  
	   inputs:
	   
	   x    vector of points to compute kernel density at for plotting
	   data vector of data points used to compute kernel density
	  
	   implicit global inputs

	   kerneltyp: type of kernel density to use in the density estimation

	              epanechnikov
		      gaussian
	   
	   */

       load qmc,smc;  

       data=qmc[.,2];
       x=seqa(-10,.1,161);
       nden=zeros(161,1);
       sig=meanc(smc[.,2]);

       i=1; do until i>rows(x);
          nden[i]=exp(-((x[i]-meanc(data))^2)/(2*sig*sig))/(sqrt(2*pi)*sig);
       i=i+1; endo;

       kde=kden(x,data);

       library pgraph;
       graphset;
       _pdate="";
       _pmsgstr="";
       _pmsgctl=0;

#IFUNIX

      let v = 100 100 640 480 0 0 1 6 15 0 0 2 2;
      wxy = WinOpenPQG(v,"XY Plot","XY");
      call WinSetActive(wxy);

#ENDIF

_pmsgstr="Number of obs   "$+ftos(rows(data),"lf",6,0)$+"\000"$+
         "Mean            "$+ftos(meanc(data),"lf",6,3)$+"\000"$+
         "Std deviation   "$+ftos(stdc(data),"lf",6,3)$+"\000"$+
         "Maximum value   "$+ftos(maxc(data),"lf",6,3)$+"\000"$+
         "Minimum value   "$+ftos(minc(data),"lf",6,3);
_pmsgctl=(.5~5.4~.15~0~2~4~0)|
         (.5~5.2~.15~0~2~4~0)|
         (.5~5.0~.15~0~2~4~0)|
         (.5~4.8~.15~0~2~4~0)|
         (.5~4.6~.15~0~2~4~0);
   _pline=(1~3~-2.5~0~-2.5~.5~1~4~3);
   _plegstr="Finite sample distribution\000Normal Approximation";
   _plegctl=2~4~.5~3.8;
   
      title("Full information semiparametric 2-step estimator for q]2[\LTrue value of q]2[=-2.5, 500 monte carlo draws");
      xy(x,kde~nden);

#IFUNIX
 
 call WinSetActive(1);
  
#ENDIF


       data=qmc[.,1];
       x=seqa(-10,.1,161);
       nden=zeros(161,1);
       sig=meanc(smc[.,1]);

       i=1; do until i>rows(x);
          nden[i]=exp(-((x[i]-meanc(data))^2)/(2*sig*sig))/(sqrt(2*pi)*sig);
       i=i+1; endo;

       kde=kden(x,data);

       library pgraph;
       _pdate="";
       _pmsgstr="";
       _pmsgctl=0;

#IFUNIX

      let v = 100 100 640 480 0 0 1 6 15 0 0 2 2;
      wxy = WinOpenPQG(v,"XY Plot","XY");
      call WinSetActive(wxy);

#ENDIF

_pmsgstr="Number of obs   "$+ftos(rows(data),"lf",6,0)$+"\000"$+
         "Mean            "$+ftos(meanc(data),"lf",6,3)$+"\000"$+
         "Std deviation   "$+ftos(stdc(data),"lf",6,3)$+"\000"$+
         "Maximum value   "$+ftos(maxc(data),"lf",6,3)$+"\000"$+
         "Minimum value   "$+ftos(minc(data),"lf",6,3);
_pmsgctl=(.5~5.4~.15~0~2~4~0)|
         (.5~5.2~.15~0~2~4~0)|
         (.5~5.0~.15~0~2~4~0)|
         (.5~4.8~.15~0~2~4~0)|
         (.5~4.6~.15~0~2~4~0);
   _pline=(1~3~-0.5~0~-0.5~.5~1~4~3);
   _plegstr="Finite sample distribution\000Normal Approximation";
   _plegctl=2~4~.5~3.8;
   
      title("Full information semiparametric 2-step estimator for q]1[\LTrue value of q]1[=-0.5, 500 monte carlo draws");
      xy(x,kde~nden);

#IFUNIX
 
 call WinSetActive(1);
  
#ENDIF