数据库模糊查询结果自动排序方法-JournalofNortheasternUniversity.PDFVIP

第29 卷第7期 东 北 大 学 学 报 ( 自 然 科 学 版 ) V ol29, No. 7 2 00 8 年 7 月 Journal of Nort heastern U n vers ty( N atural Sc ence) Jul. 2 0 0 8 孟祥福, 马宗民, 严 丽 ( , 110004) : , , , P IR , , DPR , : ; ; PIR ; : T P 3 11 : A : 1005- 3026( 200 8) 07-0960- 05 Automated Ranking of Database Fuzzy Query Results MEN G X iang-f u, MA Zong-min, YAN L i ( School of Infor mat on Sc ence Eng neer ng , N ortheastern U n vers ty, Shenyang 110004 , Ch na. Correspondent : M ENG X ang- fu , E- ma l: marx @ 126. com) Abstract: T here m ay be t oo m any tuples n an answ er after fuzzy query based on relat onal dat abase. So, t s necessary to rank query results n accordance to u ser s needs. T he rank ng process s as follow s. A ccord ng t o t he deg ree of membersh p betw een tuple and fuzzy query , t he tuples w t h d ff erent deg ree of membersh p n query results are d ff erent ated from each ot her t o form several membersh p classes. T hen, t he mproved P IR ( probab l t y nformat on ret r eval ) model and dat abase w orkload are used to analyze the degree of correlat on betw een the att r bute v alues spec f ed and unspec f ed n quer es, thus rank ng t he t uples w t h the same deg ree of membersh p due t o users preferences. As a result, the DPR ( degree of mem bersh p and probab l ty rank ng) process s proposed as an aut omat ed rank ng approach for fuzzy query results , w h ch has been proved to mprove g reat ly t he rank ng qual t y of fuzzy query result s t hrough exper me