#### DMCA

## Aggregation in Probabilistic Databases via Knowledge Compilation

### Cached

### Download Links

Citations: | 8 - 3 self |

### Citations

110 | MCDB: a monte carlo approach to managing uncertain data.
- Jampani, Xu, et al.
- 2008
(Show Context)
Citation Context ...ave considered restricted instances of the problem: they focus on aggregates over one probabilistic table of restricted expressiveness [4, 20, 16], or rely on expected values and Monte-Carlo sampling =-=[10, 12, 22]-=-. Expected values can lead to unintuitive query answers, for instance when data values and their probabilities follow skewed and non-aligned distributions [19]. Abiteboul et al. investigate XML querie... |

65 | Conditioning Probabilistic Databases.
- Koch, Olteanu
- 2008
(Show Context)
Citation Context ...luation of relational algebra queries (without aggregates) on probabilistic c-tables [18]. It has been shown that more complex tasks, such as conditioning probabilistic databases on given constraints =-=[14]-=- and sensitivity analysis and explanation of query results [11], can benefit from decomposition trees. Example 1. Figure 1 shows six pvc-tables, amongst them the suppliers table S, the products tables... |

59 |
Probabilistic Databases
- Suciu, Olteanu, et al.
- 2011
(Show Context)
Citation Context ...abilistic databases are useful to represent and query imprecise and uncertain data, such as data acquired through measurements, integrated from multiple sources, or produced by information extraction =-=[21]-=-. In this paper, we use a representation system for probabilistic data called pvc-tables. It is based on the algebraic structures of semiring and semimodule to support a mixed representation of aggreg... |

54 |
Probability: an introduction.
- Grimmett, Welsh
- 1986
(Show Context)
Citation Context ...y space (Ω, 2 Ω , Pr) that we call the probability space induced by X.The probability distribution of the sum of two independent random variables is the convolution of their individual distributions =-=[8]-=-. For instance, given two random variables x, y over positive integers, the probability that the sum of the random variables equals to 4 is the sum of the probabilities of x being 0 and y being 4, of ... |

41 | Using OBDDs for efficient query evaluation on probabilistic databases
- Olteanu, Huang
- 2008
(Show Context)
Citation Context ...r queries in 9.1, first consider the query Q ′ = π Āσψ(Q1 × · · · × Qn) which is by assumption hierarchical. It follows that, given a tuple t ∈ Q ′ , its annotation Φ = ∑ φi is a read-once expression =-=[17, 21]-=-; moreover, Φ can be compiled into a d-tree whose size is bounded by the number of its variables. By Propositions 2 and 3, computing the probability distribution of such a d-tree – and thus the probab... |

37 | Provenance for aggregate queries. - Amsterdamer, Deutch, et al. - 2011 |

20 | The trichotomy of HAVING queries on a probabilistic database.
- Ré, Suciu
- 2009
(Show Context)
Citation Context ...sions yij ⊗ vi: x1(y11 ⊗ 10 + y12 ⊗ 50) + x2(y21 ⊗ 11 + y22 ⊗ 60) + x3(y33 ⊗ 15 + y34 ⊗ 40). ✷ Ré et al. consider queries ϖ∅;γ←AGG(C)σφ(R1 × · · · × Rn) in which π∅σφ(R1 × · · · × Rn) is hierarchical =-=[19]-=-; these are subsumed by Q hie . For such queries involving aggregation without grouping, the neutral element of the aggregation monoid may safely be in the database without jeopardising query tractabi... |

17 | Incremental query evaluation in a ring of databases.
- Koch
- 2010
(Show Context)
Citation Context ... aggregates on probabilistic data [1]. An algebra proposed by Koch represents annotations and data values as rings which enables efficient incremental view maintenance in the presence of aggregations =-=[13]-=-. Our approach considers the problem of exact probability computation for positive relational algebra queries with aggregates on pvc-tables. The core of our technique is a procedure that compiles arbi... |

16 | Probabilistic top-k and ranking-aggregate queries
- Soliman, Ilyas, et al.
(Show Context)
Citation Context ...d. Existing approaches to aggregates in probabilistic databases have considered restricted instances of the problem: they focus on aggregates over one probabilistic table of restricted expressiveness =-=[4, 20, 16]-=-, or rely on expected values and Monte-Carlo sampling [10, 12, 22]. Expected values can lead to unintuitive query answers, for instance when data values and their probabilities follow skewed and non-a... |

14 | Sensitivity analysis and explanations for robust query evaluation in probabilistic databases.
- Kanagal, Li, et al.
- 2011
(Show Context)
Citation Context ...robabilistic c-tables [18]. It has been shown that more complex tasks, such as conditioning probabilistic databases on given constraints [14] and sensitivity analysis and explanation of query results =-=[11]-=-, can benefit from decomposition trees. Example 1. Figure 1 shows six pvc-tables, amongst them the suppliers table S, the products tables P1 and P2, and the table P S pairing suppliers and products. T... |

14 | PIP: A database system for great and small expectations.
- Kennedy, Koch
- 2010
(Show Context)
Citation Context ...ave considered restricted instances of the problem: they focus on aggregates over one probabilistic table of restricted expressiveness [4, 20, 16], or rely on expected values and Monte-Carlo sampling =-=[10, 12, 22]-=-. Expected values can lead to unintuitive query answers, for instance when data values and their probabilities follow skewed and non-aligned distributions [19]. Abiteboul et al. investigate XML querie... |

14 | Aggregate queries over conditional tables
- Lechtenbörger, Shu, et al.
(Show Context)
Citation Context ...ries with aggregates can be represented as pvc-tables of polynomial size. This contrasts with main-stream representation systems such as pc-tables [21], which can require an exponential-size overhead =-=[15]-=-. The problem of query evaluation is #P-hard already for simple conjunctive queries [21]. Aggregates are a further source of computational complexity: for example, already deciding whether there is a ... |

5 |
An improvement on the complexity of factoring read-once Boolean functions
- Golumbic, Mintz, et al.
(Show Context)
Citation Context ...cognise read-once expressions, i.e., expressions where each variable occurs once, and hence factorise expressions based on algebraic rewritings such as the associativity and commutativity laws, e.g., =-=[6, 18]-=-. In particular, this approach allows to factor expressions into complex subexpressions and not only into one variable and the residual. The last rule decomposes Φ into sub-expressions Φs, for each s ... |

5 | Querying uncertain data with aggregate constraints - Yang, Wang, et al. - 2011 |

1 |
Evaluation of Aggregate Queries in Probabilistic Databases
- Han
- 2011
(Show Context)
Citation Context ...are discussed. Besides exact computation, decomposition trees also allow for approximate probability computation [18]. Due to lack of space, we refer the reader to the MSc thesis of the second author =-=[9]-=-. The pvc-tables can be extended to cope with continuous probability distributions, similar to the extensions of pc-tables in the PIP system [12]. 2. PRELIMINARIES 2.1 Induced Discrete Probability Spa... |