Research On Algorithm For Attribute Reduct & Core Computation Based On Rough Set Theory

Rough Set theory is a mathematic tool for analyzing the different kinds ofinformation with no consistency, accuracy and integrity. At present, it has been widelyapplied in Artificial Intelligence, Pattern Recognition, and Knowledge Discovery andso on. Attribute Reduction is the core content of the Rough Set theory research. It is aprocess for deleting the redundant attribute of the system on condition that it keeps theclassification ability of the Decision-making System unchanged and the improvedsystem keeps the function of the original Decision-making System. The simplified ruledatabase can be formed after the Attribute Reduction so that we can make quick andaccurate decisions.In this thesis, two algorithms on the basis of the original Rough Set AttributeReduction research are presented as following: the Calculation of Core AttributeAlgorithm based on the Importance of Attribute and the Reduction Algorithm ofAttribute based on Importance of Attribute. The Core Attribute is not delectable in theinformation system, which is the intersection of all the Reduction Algorithms ofAttribute. This thesis gets the Core Attribute by using the Importance of Attribute tocalculate the importance of the separated attribute in the Decision Table firstly. Whenimportance of a certain attribute is more than zero, and this attribute will be the CoreAttribute, incorporating this attribute into the set of the Core Attribute and getting theset of the Core Attribute in the Decision Table at last with a round of loop computingjudgment. The Reduction Algorithm of Attribute is based on Importance of Attributeby using the Calculation of Core Attribute Algorithm to get the Core Attribute. At first,to get the Core Attribute by using the former Arithmetic; Secondly, to calculate theother importance of the attributes in the condition attribute except the core on basis ofthe Core Attribute, then add the biggest importance of the attribute into the core set toform the new set RED. Thirdly, to circulate respectively on the basis of RED until allthe importance of the rest attributes become zero, from which the REDn is got, as isAttribute Reduction.According to the comparative analysis of the experiment, the improved calculation has higher efficiency and can make sure to get the Minimum Reduction.