we have been witnessing much interest in uncertain data management in many application areas such as data cleaning , sensor networks , information extraction , etc. Much research effort has been devoted to several aspects of uncertain data management, including data modeling , skyline queries, top-k queries, nearest neighbor search , spatial queries , XML documents , etc. There has been some work dealing with aggregate query processing over uncertain data. Some of them were devoted to developing efficient algorithms for returning the expected value of aggregate values . For example in the authors study the problem of computing aggregate operators on probabilistic data in an I/O efficient manner. With the expected value semantics, the evaluation of SUM queries is not very challenging. ALL_SUM, but they only propose an efficient approach for computing the expected value. Approximate algorithms have been proposed for probabilistic aggregate queries. The Central Limit theorem an be used to approximately estimate the distribution of sums for sufficiently large numbers of probabilistic values SUM aggr queries are critical for many applications that need to deal with uncertain data. In this paper, we addressed the problem of evaluating ALL_SUM queries. After proposing a new recursive approach, we developed an algorithm, called Q_PSUM, which is polynomial in the number of SUM results. Then, we proposed a more efficient algorithm, called DP_PSUM, which is very efficient in the cases where the aggr attribute values are small integers or real numbers with small precision. We validated our algorithms through implementation and experimentation over synthetic and real-world data sets.
You are here: Home / ieee projects 2013-2014 / Supporting real attribute values with small precisions