சுருக்கம்

A Framework Outline for Word Catalog Based Model in Cloud

Mr.C.Sivakumar, Mr.M.Arulprakash M.Tech

Ranked search greatly enhances system usability by enabling search result relevance ranking instead of sending undifferentiated results, and further ensures the file retrieval accuracy. We add a ranking procedure that turns a pairwise preference relation into a complete preorder of the option The statistical measure approach, i.e., relevance score, from information retrieval to build a secure searchable index, and develop a one-to-many order-preserving mapping technique to properly protect those sensitive score information. The resulting design is able to facilitate efficient server-side ranking without losing keyword privacy. Thorough analysis shows that our proposed solution enjoys “as strong-aspossible” security guarantee compared to previous searchable encryption schemes, while correctly realizing the goal of ranked keyword search. We introduce a general preference based racing (PBR) algorithm that provides the basic statistical information’s which is needed to solve the selection problem, notably estimates of the yi,j and corresponding confidence intervals. It contains a subroutine that implements sampling strategies for the different ranking models based on user preferences. A Manipulation mining system will take care of calculating the users interest and provide an optimistic approach to fetch the data based on user’s interest.