Open AccessReduction algorithm based on finding the maximum mutual information in incomplete information systems
Author(s) -
Feng Wang,
Manting Zhang
Publication year - 2019
Publication title -
journal of physics. conference series
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.21
H-Index - 85
eISSN - 1742-6596
pISSN - 1742-6588
DOI - 10.1088/1742-6596/1237/2/022020
Subject(s) - mutual information , attribute domain , rough set , mathematics , reduction (mathematics) , entropy (arrow of time) , interaction information , data mining , algorithm , computer science , mathematical optimization , statistics , physics , geometry , quantum mechanics
Aiming at such drawbacks of the existing incomplete information system attribute reduction algorithms as high time complexity and insufficiency of the reduction result completeness, this paper proposes an attribute reduction algorithm based on finding the maximum mutual information via weakening the equivalence class in classical rough set theory into tolerance class and through combining the definition of information entropy with the tolerance class. Taking the mutual information of condition attribute and decision attribute as the iterative criterion, and the empty set as the initial reduction, the algorithm superposes the condition attribute set corresponding to the maximum value of the mutual information of condition attributes and decision attributes, thus equating the mutual information of condition attribute set and decision attribute with the information entropy of decision attribute, obtaining the relative reduction of the incomplete decision system and guaranteeing the completeness of the reduction result. In the solution of the attribute set tolerance class, an algorithm with lower time complexity is adopted, which can reduce the time complexity of the whole algorithm. The feasibility of the algorithm is illustrated by an example.