On 10.04.2021, at 08:10, Antoine Monmayrant <antoine.monmayr...@laas.fr> wrote:
>
>
> On 09/04/2021 23:55, Heinz Nabielek wrote:
>> xfpoly does generally a good job for me, sometimes I would wish that the
>> filling colour could be made transparent.
> This is a much needed improvement of the scilab graphical stack that
> currently does not provide any transparency.
> I think this is not an easy improvement.
> At the moment, my workaround is to plot everything I need, export as svg and
> than add the transparency I need in the svg using inkscape or directly
> editing the svg file with a text editor...
>> Heinz
>>
>> PS 1: Is there a new version of the 2011 BetweenCurves around?
> Er, no, it was just a dirty hack I needed for my own publications and I think
> 2011 is the most recent one.
> I can try to see how to improve it if this can improve scilab...
I had initiated by log vs lin plot with
>> plot2d([0 80],[1 100], style=0,logflag = "nl");
but BetweenCurves starts with its own plot and here I would not know, what to
do...
Heinz
BTW, how do the French infection rates look like?
>>
>> PS 2: Any suggestion to make my clumsy coding more elegent, is highly welcome
>>
>> PS 2: BTW, since the recent lockdown, infection rates are coming down in
>> Austria.....
>>
>>
>> A=[12.628 13.942 17.077 17.054 15.594 14.976 14.796
>> 11.875 13.448 16.504 17.717 19.447 16.099 13.302 13.762
>> 16.032 19.492 22.098 20.425 21.087 20.649 14.268
>> 19.402 22.525 26.862 23.514 27.603 23.851 15.830 21.570
>> 28.682 26.109 29.974 28.727 24.705 21.458 27.087
>> 28.401 33.670 35.119 33.962 28.120 21.301 27.244 37.467
>> 37.715 39.490 37.569 30.480 27.098 38.366 36.951
>> 35.097 43.759 39.299]';
>> d=(1:length(A))';
>> up=10^(d/53);
>> M=[ones(up) up];
>> aa=M\A;
>> B=inv(M'*M);
>> DD=(1:110)';
>> U=10^(DD/53);
>> MM=[ones(U) U];
>> yh = M*aa; //Fitted values yh to approximate measured y's
>> e=A-yh; //Errors or residuals
>> SSE=e'*e; //Sum of squared errors
>> ybar=mean(A); R2=1-SSE/sum((A-ybar)^2);
>> [m n]=size(M);
>> MSE = SSE/(m-n-1); //Mean square error
>> C=MSE*B // covariance matrix
>> sig=sqrt(MSE);
>> seb=sqrt(diag(C));
>> [aa seb]
>> [n pp]=size(M);
>> CONF=.95; alpha=1-CONF;
>> ta2 = cdft('T',n-pp,1-alpha/2,alpha/2); //t-value for alpha/2
>> yhh= MM*aa;
>> p=sig*sqrt(diag(1+MM*B*MM'));
>> N=[yhh+ta2*p yhh-ta2*p];
>> polyX = [DD;flipdim(DD,1)];
>> polyY = [N(:,1);flipdim(N(:,2),1)];
>> plot2d([0 80],[1 100], style=0,logflag = "nl");
>> xgrid;
>> xfpoly(polyX, polyY,6);
>> plot(DD,MM*aa,'g.-');
>> plot(d,A,'b.') ;
>> title('AUSTRIA daily infection rates per 100,000','fontsize',5);
>> xlabel('days since 1 Feb 2021','fontsize',3);
>> ylabel('number of infections per day per 100,000','fontsize',3);
>> legend('data from Johns Hopkins GitHub','95% confidence range','model
>> prediction','AUSTRIA recorded',4);
>>
>>
>>
> _______________________________________________
> users mailing list
> users@lists.scilab.org
> http://lists.scilab.org/mailman/listinfo/users
_______________________________________________
users mailing list
users@lists.scilab.org
http://lists.scilab.org/mailman/listinfo/users
