evalu queri result signific databas random permut test studi classifi perform random techniqu graph mine graph data activ research area data mine method algorithm propos identifi structur graph evalu result lack framework statist hypothesi test focu paper random techniqu unweight undirect graph random import approach assess statist signific data mine result input graph random method sampl data class graph share structur properti input graph describ altern algorithm base local edg swap metropoli sampl test framework graph data set mine algorithm applic graph cluster frequent subgraph mine queri signific databas random guest editor introduct special issu mine learn graph featur select taxonomi applic paleontolog band structur binari matric approxim ratio biclust problem biclust consist simultan cluster row column matrix submatric induc pair row column cluster uniform paper approxim optim biclust appli cluster algorithm independ row column input matrix show solut yield worst case approxim ratio norm valu matric norm real valu matric keyword approxim algorithm biclust cluster close set label data abstract close set successfulli appli context compact data represent associ rule learn descript paper show label data close set adapt predict discrimin purpos conveni contrast cover properti posit neg exampl formal justifi set character space relev combin featur discrimin target class practic identifi relev irrelev combin featur close set applic appli compact emerg pattern essenti rule learn descript subgroup discoveri character implic inject partial order cost learn direct cut classifi vertic digraph import machin learn set applic learn problem digraph characterist properti target concept correspond direct cut ii total cost find cut bound priori iii target concept chang due hidden context motiv classifi intermedi product process manufactur car control flow softwar faulti correct process repres digraph concept monoton typic fault intermedi product present stage product concept depend hidden variabl pre assembl part vari fault occur charg order trade cost faulti product cost need find fault tight perform guarante find bug need perform guarante propos machin learn literatur distinguish character implic inject partial order cross dmine binari numer attribut problem relat itemset mine binari attribut data set numer attribut data biogeograph data numer attribut correspond environment variabl binari attribut encod presenc absenc speci environ viewpoint itemset mine task select small collect interest itemset numer attribut viewpoint numer attribut task constrain search local pattern cluster binari attribut give formal definit problem discuss theoret give simpl constant factor approxim algorithm show experi biogeograph data algorithm captur interest pattern found itemset mine cluster mine close set multi drelat data investig problem mine close set multi relat databas previou work introduc semant algorithm mine close set multi relat databas insight implic semant choic relationship lack investig show semant choic import impli properti turn affect rang algorithm mine set interest question semin lcm algorithm uno al upgrad multi relat problem lcm attract run time linear number close set store output order avoid duplic provid posit answer question semant choic report experi evalu scalabl applic upgrad algorithm benchmark problem cost learn direct cut classifi vertic digraph import machin learn set applic learn problem digraph characterist properti target concept correspond direct cut ii total cost find cut bound priori iii target concept chang due hidden context motiv classifi intermedi product process manufactur car control flow softwar faulti correct process repres digraph concept monoton typic fault intermedi product present stage product concept depend hidden variabl pre assembl part vari fault occur charg order trade cost faulti product cost need find fault tight perform guarante find bug need perform guarante propos machin learn literatur distinguish approxim ratio biclust problem biclust consist simultan cluster row column matrix submatric induc pair row column cluster uniform paper approxim optim biclust appli cluster algorithm independ row column input matrix show solut yield worst case approxim ratio norm valu matric norm real valu matric keyword approxim algorithm biclust cluster horn axiomat sequenti data close set label data abstract close set successfulli appli context compact data represent associ rule learn descript paper show label data close set adapt predict discrimin purpos conveni contrast cover properti posit neg exampl formal justifi set character space relev combin featur discrimin target class practic identifi relev irrelev combin featur close set applic appli compact emerg pattern essenti rule learn descript subgroup discoveri horn axiomat sequenti data summar sequenti data close partial order coproduct transform lattic close partial order sampl strategi find frequent set discov unbound episod sequenti data dfirst strategi find frequent set associ rule discoveri problem consist identifi frequent itemset databas form condit implic rule algorithm difficult part task find frequent set exist wealth algorithm problem variat case gener recent fulli approach algorithm breadthfirst search depth search lattic itemset paper propos develop search strategi 