Parallel Attribute Reduction Algorithm for Complex Heterogeneous Data Using MapReduce

Algorithm 4

Single condition attribute subset neighborhood-Reduce.

Input: <KEY_M, VALUE_M>

Output: <KEY_R, VALUE_R>

   // let KEY_R be the set of different key, and VALUE_R be the set of sample IDs subset value' with the same key'.

 begin

  <KEY_R, VALUE_R>=

  for <key, value>in <KEY_M, VALUE_M>do

   if key is not appeared in <KEY_R, VALUE_R>

    <key, value>=<key, value>

   else

   if key=key_k

    <KEY_R, VALUE_R>=<KEY_R, VALUE_R>-<key, value>

    value_k=value_k value // combine samples with the same key

   end if

   end if

   <KEY_R, VALUE_R>=<KEY_R, VALUE_R> <key, value>

  end for

 end