Against the background of Internet, Internet of Things (IoT) and Big Data, it has become necessary to solve the problem of data uncertainty. It’s not appropriate to manage data in the traditional database due to the inaccurate measurements, unideal measuring conditions, restricted budget or man-caused blurred private information and so on. In that, uncertainty database has been investigated adequately, based on which, the Top-k query has particularly been discussed widely and is the hot topic in the academy and industry.The thesis reviewed the Uncertainty-Lineage Database(ULDB) and the some Top-k queries based on uncertain database that has been proposed already. Then the thesis proposed two new Top-k queries based on uncertainty database, OS-UTop-k and OI-UTop-k. The most important significance of them is that they regard the object(x-tuple) as the query target and get the results with more practical value. The result of the present Top-k queries based on uncertain database is all based on instance-level (tuple-level), which is more friendly to the computers rather than the users. The new queries processing method proposed by the thesis, strived to diminish the impact of efficiency due to changing query target. At the last, some experiments proved the feasibility of the algorithm proposed by the thesis. |