wolffd@0: function [bnt_data, old_values] = transform_data_into_bnt_format(data,cnodes)
wolffd@0: % TRANSFORM_DATA_TO_BNT_FORMAT Ensures discrete variables have values 1,2,..,k
wolffd@0: % e.g., if the values of a discrete are [0 1 6], they must be mapped to [1 2 3]
wolffd@0: %
wolffd@0: % data(i,j) is the value for i-th node in j-th case.
wolffd@0: % bnt_data(i,j) is the new value.
wolffd@0: % old_values{i} are the original values for node i.
wolffd@0: % cnodes is the list of all continous nodes, e.g. [3 5] means the 3rd and 5th node is continuous
wolffd@0: %
wolffd@0: % Author: yimin.zhang@intel.com
wolffd@0: % Last updated: Jan. 22, 2002 by Kevin Murphy.
wolffd@0: 
wolffd@0: num_nodes=size(data,1);
wolffd@0: num_cases=size(data,2);
wolffd@0: old_values=cell(1,num_nodes);
wolffd@0: 
wolffd@0: for i=1:num_nodes
wolffd@0:   if (myismember(i,cnodes)==1)  %cts nodes no need to be transformed 
wolffd@0:     %just copy the data
wolffd@0:     bnt_data(i,:)=data(i,:);
wolffd@0:     continue;
wolffd@0:   end
wolffd@0:   values = data(i,:);
wolffd@0:   sort_v = sort(values); 
wolffd@0:   %remove the duplicate values in sort_v
wolffd@0:   v_set = unique(sort_v);  
wolffd@0:   
wolffd@0:   %transform the values
wolffd@0:   for j=1:size(values,2)
wolffd@0:     index = binary_search(v_set,values(j));
wolffd@0:     if (index==-1)
wolffd@0:       fprintf('value not found in tranforming data to bnt format.\n');   
wolffd@0:       return;
wolffd@0:     end
wolffd@0:     bnt_data(i,j)=index;
wolffd@0:   end
wolffd@0:   old_values{i}=v_set;
wolffd@0: end
wolffd@0: 
wolffd@0: 
wolffd@0: %%%%%%%%%%%%
wolffd@0: 
wolffd@0: function index=binary_search(vector, value)
wolffd@0: % BI_SEARCH do binary search for value in the vector
wolffd@0: % Author: yimin.zhang@intel.com
wolffd@0: % Last updated: Jan. 19, 2002
wolffd@0: 
wolffd@0: begin_index=1;
wolffd@0: end_index=size(vector,2); 
wolffd@0: index=-1;
wolffd@0: while (begin_index<=end_index)
wolffd@0:   mid=floor((begin_index+end_index)/2);
wolffd@0:   if (isstr(vector(mid)))
wolffd@0:     % need to write a strcmp to return three result (< = >)
wolffd@0:   else
wolffd@0:     if (value==vector(mid))
wolffd@0:       index=mid;
wolffd@0:       return;
wolffd@0:     elseif (value>vector(mid))
wolffd@0:       begin_index=mid+1;    
wolffd@0:     else
wolffd@0:       end_index=mid-1;
wolffd@0:     end
wolffd@0:   end
wolffd@0: end
wolffd@0: return;