Threshold-based probabilistic top-k dominating queries

Abstract

Recently, due to intrinsic characteristics in many underlying data sets, a number of probabilistic queries on uncertain data have been investigated. Topk dominating queries are very important in many applications including decision making in a multidimensional space. In this paper, we study the problem of efficiently computing top-k dominating queries on uncertain data. We first formally define the problem. Then, we develop an efficient, threshold-based algorithm to compute the exact solution. To overcome some inherent computational deficiency in an exact computation, we develop an efficient randomized algorithm with an accuracy guarantee. Our extensive experiments demonstrate that both algorithms are quite efficient, while the randomized algorithm is quite scalable against data set sizes, object areas, k values, etc. The randomized algorithm is also highly accurate in practice.

Best position algorithms for efficient top-k query processing

“…Top-k queries have attracted much interest in many different areas such as network and system monitoring [8,23], information retrieval [18,22,32,36], sensor networks [34,39], multimedia databases [9,17,31], spatial data analysis [10], P2P systems [3,4], data stream management systems [13,21,27,30], probabilistic databases [33,35,40], temporal databases [26], etc. The main reason for such interest is that they avoid overwhelming the user with large numbers of uninteresting answers, which are resource-consuming.…”

Promotion Analysis Based on Fitting in Multi-Level Space

“…Probabilistic formulations for top-k and ranking aggregate queries is introduced in probabilistic databases [13]. Threshold-based algorithm is developed to compute the exact solution [14]. We compute measure based on fitting for fast processing due to the measures sorted according to rank of the object in monotone decreasing order.…”

Top-<italic>k</italic> Dominating Queries on Incomplete Data

“…Intuitively, a TKD query could identify the k most popular movies for moviegoers. Because of its large application base, the TKD query has received lots of attention from the database community [3], [4], [5], [6], [7], [8], [9]. Nonetheless, we would like to highlight that existing work related to this query only focuses on complete data or uncertain data.…”

“…Although the TKD query over complete data or uncertain data has been well studied, TKD query processing on incomplete data still remains a big challenge. This is because existing techniques [3], [4], [5], [6], [7], [8] cannot be applied to handle the TKD query over incomplete data efficiently. Specifically, the R-tree/aRtree and the transitivity of dominance relationship used in traditional and uncertain databases are not directly applicable to incomplete data.…”