Parallel Attribute Reduction Algorithm for Complex Heterogeneous Data Using MapReduce

Algorithm 3

Single condition attribute subset neighborhood-Map.

Input: Single condition attribute subset, C; the hash bucket B; and a data split S_i

Output: <KEY_M, VALUE_M>

   //key represents whether the sample belongs to the positive region; let key of the sample in positive region be 1, and key of the sample in boundary be 0; value represents all samples’ ID that have the same key

 begin

  <KEY_M, VALUE_M>=

  for each do

   let key_i=0 //assuming that this sample does not belongs to the positive region under C

   let value=the ID of and given that

   for each //traversing the hash bucket where a neighborhood probably exists

    if is the neighborhood of , but they have different decision attribute values

     let key_i=1

    else

     break

    end if

   end for

   <KEY_M, VALUE_M>=<KEY_M, VALUE_M> <key_i, value_i>

  end for

 end