Research Article

Parallel Attribute Reduction Algorithm for Complex Heterogeneous Data Using MapReduce

Algorithm 1

Hash-Map function.
Input: condition attribute subset, C; a data split Si
Output: <KEYHM, VALUEHM> // let KEYHM be the set of hash value of each sample, and VALUEHM be the set of sample ID
begin
  <KEYHM, VALUEHM>=
  for each do
   let key=hash(xi);
    // =key, where x0 is a special sample in universe U, which is satisfied with , and f is the information function.
   let value=the ID of
   <KEYHM, VALUEHM>=<KEYHM, VALUEHM>= <key, value>
  end for
end