The Study Of Rough Data Querying Based On RRDB

Rough set theory, introduced by Pawlak in 1982, which is a technique for managing the uncertainty and imperfection, can analyze incomplete information effectively. It builds on the basis of the classification mechanism, and applies the knowledge base to describe the uncertain or imprecise knowledge. The most notable difference between rough set theory and other theory which can deal with the problem of uncertainty and imprecision is that it does not need any apriori information of the problem's data.Beaubouef. T incorporated the rough set theory into the classical RDM (Relational Database Model), and suggested the term of "Rough Relational Database Model (RRDM)" to describe the uncertain information system. The RRDM has several features in common with the classical RDM, and the primary difference between them is that in RDM, an attribute's value must be atomic; and that in RRDM, an attribute can be composed of one or more atomic values. So RRDM is actually a multi-valued information system. RRDM is a combination of basic relational database theory and several rough set theory's basic concepts, such as equivalence class, the upper approximation, the lower approximation, and so on. It is the extension of standard relational database, and can enhance the information expression's ability of the standard relational database.Data querying is the core operation of database system, and data querying based on RRDM is called Rough Data Querying. Rough Data Querying is classified into three kinds: crisp querying, rough complete querying and rough combinatorial querying. At present many scholars has done some reserach on the implementation of the rough data querying, and the main idea of it's implementation is expanding the original search conditions according to the indiscernibility relationship in attribute's domain. Because the expanded search conditions need more calculation, the querying gets very slow.In this paper, we present a solution called rough data querying based on vector-encoding. Firstly encode the data of multi-valued attribute into the single-valued data according to the indiscernibility relationship in attribute's domain, and then execute the query on the encoding single-valued data despite the indiscernibility relationship to make the implementation of rough data querying simpler and more efficient. Further more, such a query is based on the standard SQL, and it can easily realize the functions of RRDM in the existing relational database, so it has a good practicability.The main works and achievements of the article are:1. It gives the methods of rough data querying based on vector-encoding, presents the detailed algorithm of vector-encoding, and discusses how to maintain the vector-value;2. It presents the concrete solution based on vector-encoding to three kinds rough data querying;3. We carry out the corresponding experiment to prove the performances of the algorithms. Based on the analysis of experimental results, we draw some useful conclusions.