Fuzzy Rough Set Attribute Reduction And Multi-fuzzy Decision Information Fusion

In Rough Set the attribute reduct is the crucial idea.but before finding attribute reducts with rough set,continuous values must be transformed into crisp values to handle,to some extent,the information loss occurs in the process of discretization. The integration of fuzzy sets and rough sets can preserve the information of continuous values attribute,with fuzzy rough sets attribute reduct can get higher accuracy than with rough sets.Using most of the fuzzy attribute algorithm,the most important fuzzy attribute subsets(reduct) could be found.because of only some of useful information being contained in this attribute reduct, the classification error accuracy couldn't be decreased any more,In order to make full use of the information provided by every fuzzy attribute reduct,in this paper,we present a novel algorithm,with this algorithm several fuzzy subset could be found.to these fuzzy subset,we present a multiple fuzzy decision tress fusion algorithm based on fuzzy integral,multiple fuzzy decision tree classifier fusion can make full use of information provided by every fuzzy decision tress, the fusion leads to an higher classification accuracy than single fuzzy decision tree.