Research Article
Rule Extraction Model Based on Decision Dependency Degree
| | Input: decision system | | | Output: the decision rule set | | (1) | Call DecDep_Red reduction model in [29] to get ; initialize rule set | | (2) | Call RadSort and RedSet (, ) in [29] to get | | (3) | for each do | | (4) | Let ; call RadSort and DecDep_Deg in [29] to get of P | | (5) | end | | (6) | Add relevant attributes into Array in descending order according to the achieved values | | (7) | for each do | | (8) | Let and call RadSort | | (9) | Call ConDup to get and | | (10) | Execute a mark procedure to get | | (11) | end | | (12) | for each do | | (13) | if , then | | (14) | For all attributes of , add attributes with original value to P, add attributes with “!” to | | (15) | end | | (16) | if , then | | (17) | ; Restore to its original value and update | | (18) | end | | (19) | for each do | | (20) | | | (21) | end | | (22) | Call RadSort and get | | (23) | while | | (24) | Restore to its original value | | (25) | Update P and | | (26) | Call RadSort and update | | (27) | end | | (28) | Replace each “!” with “” for all attribute value of ; and update | | (29) | end | | (30) | Call RedSort | | (31) | Call RedSet , delete duplicate rules, and delete rules only with “” to update of | | (32) | Call RedSort | | (33) | Call RedRul to get | | (34) | For = , update − | | (35) | Delete “” in each rule of to get |
|
