% estimation of transition probability matrix
 
function [sita,n]=esti1(n,sita)
global y my t n0 ty m lambda p np
epsi=1;            
n0=epsi*ones(my,np*my^m);                  % occurence table
uni=ones(1,my);
delta=zeros(my,1); 
delta(y(t))=1;
ty=ty+(p-1)*my^m;
n(:,ty)=lambda*(n(:,ty)+delta)+(1-lambda)*n0(:,ty);
sita(:,ty)=n(:,ty)/(uni*n(:,ty));