The Research On Query Evaluation And Top-k Query In Probabilistic Relation

It is uncertain for the data in the information retrieval, sensor data and image processing. The traditional database can not do very well with it, so the research in probabilistic database becomes more and more important.This paper analyses the probabilistic database model, and discusses the shortcoming which can not make the reasonable project operation in the probabilistic database model.Since different query methods and different operation order may be NP or errors. This paper defines different operation order for different tuples in probabilistic relations, and presents the query algorithm of probabilistic relations. The query in a single table is PTIME.If join queries is satisfied with the incompatiable project or the mutual independence project.The incompatiable project is made firstly, the mutual independence project is made secondly. If the join query may be devided into two PTIME query, the join query must be devided into two query with a join sign.This article introduces new probabilistic formulations for top-k and ranking-aggregate queries into probabilistic databases. We construct a generic processing methods to supporte two query types, and employee existing query processing and indexing capabilities in current RDBMS to implement Top-k computation. In order to compute preferably the result, the original probabilistic relation is coded. The methods encapsulates a state space model and efficient search algorithms to compute query answers, this minimizes the number of accessed tuples and the size of materialized search space to compute query answers. The experiment shows that the arithmetic is PTIME, its overload is relative with the relationship among data.