Application Of The Reductionalgorithm Based On Entropy On Top Coal Caving Capability In Steep Seam

In recent years, the rough set theory has been a hot spot. With the development ofinformation theory, information entropy has great advantages in imprecise and incompleteproblem, so it has used to be a measurement of uncertainty problem in rough set research.This is a very good development for rough set, especially in the knowledge reduction. In thisarticle, it is mainly studying the problem of knowledge reduction from the aspect of theinformation theory, and then applied it in the steep seams of top coal caving mining problems.Firstly, the basic concepts and knowledge of rough set theory, information entropy andthe steep coal seam sub-level caving mining are related and analyzed. Because informationexpression and algebra expression of knowledge reduction is consistent, knowledge reductioncould be studied from the aspect of the information theory, so it is to the benefit of solveinguncertainty problem.Secondly, adopt a new kind of information entropy definition and put forward itsimproved algorithm. A new information entropy definition is applied to the CEBARKNCalgorithm based on conditions information entropy and MIBARK algorithm based on mutualinformation entropy. So we find that the new algorithm reduces calculation and improvesoperation efficiency. Analyzed the two improved algorithms and study their suitable scoperespectively.Thirdly, the improved CEBARKNC algorithm based on conditions information entropyis used to study of top coal caving characteristics in steep seam. Through the establishment ofreduction model，many technological factors and geological factors are reduced by the newalgorithm to get the simplest reduction, analyze the result. Acorrding to the reduction there becan generate rules. It is benefited of make the right judgment and decision for top coal cavingcapability in steep seam.There are many attribute reduction algorithm, but not any algorithm can get the best results on all the problems. So, if we want to get a best reduction, it is still a NP-hard problem.As new effective theory, rough set and information entropy should mix together in order toreduce attribute and have get better reduction results.