Researching For Uncertain Database And Related Theoretical Issues

With the development of data acquisition and processing techniques, there is increasing demand for processing uncertain data. Uncertain data can be classified into two categories, one is that the data itself is uncertain, such as sensor networks, radio frequency identification data(RFID), Internet data, such as location-based data; the other is that the data itself is determined, but for practical needs, processed these data into uncertain data artificially, such as data k-anonymity privacy model. Traditional relational posed a great challenge in data model, data query when uncertain data appears.The main research of this paper are as follows:Defined horizontal FDs and vertical FDs on Attribute-or data model that our decomposition will relay on them, and propose a horizontal partitioning algorithm, which is able to generate the uncertain relation decomposition conditions and solve the problem of low degree of decomposition.Presented decomposition method for the k-anonymous data. reduce the data size of k-anonymity relations, improve query efficiency and data availability.Proposed three type FDs in uncertain relation. Gave the normal forms definition of uncertainty relation, which can be used in standardized designing of uncertain relation.