An Incremental Algorithm for Mining Default Definite Decision Rules from Incomplete Decision Tables

The present paper puts forward an incremental algorithm for extracting default definite rules proposed by us from incomplete decision table using semi-equivalence classes derived from a semi-equivalence relation and their meet and join blocks on the universe. After default definite decision rules and constraint rules are acquired from the incomplete decision table, the incremental algorithm is used to modify them when new data is added to the incomplete information table. It does not need to process the original dataset repeatedly but only updates related data and rules. So it is effective in performing mining tasks from incomplete decision table. Through an example, a procedure for mining and revising rules is illustrated.