Research Article

A Parallel Attribute Reduction Method Based on Classification

Algorithm 1

Positive region acquisition algorithm based on classification.
(i)Input: Decision table
(ii)Output:Positive region of the decision table.
(iii)Step 1: let, domain, .
(iv)Step 2: Initialize the attribute index arrayand instance index arrayof the domain in the decision table, initialize a new domain as, add all instances in the new domain to the initial domain.
(v)Step 2.1 for each in, classify each attribute in an order.
 Step 2.2 for each in, classify the according to the attribute.
 Step 2.3 for eachin, insert it into the corresponding according to its corresponding attribute value.
(vi)Step 2.4 () goto Step 2.1
(vii)Step 2.5; (i<) goto Step 2.2
(viii)Step 3: for each, traverse all the last.
(ix)Step 4: ifgoto Step 6
(x)Step 5: goto Step 3
(xi)Step 6: return