A heuristic reduction algorithm based on entropy of attribute

作者：Xinying Chen^*; Yuefan Liu

来源：Journal of Convergence Information Technology, 2011, 6(3): 209-216.

DOI：10.4156/jcit.vol6.issue3.24

• Attribute reduction
• Bottom-up design
• Data sets
• Distinguishable Relation
• Distributed Control System
• Fault diagnosis expert system
• Heuristic information
• Reduction algorithms
• Rough set

When the data sets get larger at both dimensions and volumes, the classical reduction algorithm will turn out to be extremely unpractical. The paper tries to find a more efficient heuristic reduction algorithm. It defines the "distinguishable relation of attribute set"and "entropy of attribute set "concepts based on Rough Set theory. Against the new concepts, also proposes a new algorithm "heuristic reduction algorithm based on entropy of attribute". The heuristic algorithm adopts the bottom-up design, and gets attribute reduction with the heuristic information-entropy of attribute. The complexity of the algorithm in space is O(m), m=|A|. The theoretical analysis and experimental results of UCI datasets show that the way proposed here is feasible and meaningful, and provides a theoretical method for the parameter adjustment in networked fault diagnosis expert system and distributed control system of dairy enterprises.

• 出版日期2011