323,004 research outputs found
Is Indeed Fish the Best Choice in Rimini?
No full textPreferences depend on the context. For instance, given that it is at seaside, we usually prefer fish to meat for a dinner in Rimini, while probably preferences would be different if we are in Bressanone. On the other hand, Rimini is in Romagna and so the famous “Piadina romagnola” would also be a good choice. In a scenario like this, in which preferences are defined in different contexts, we address the problem of inferring which are the preferences that should be used for answering queries in a given context. To this end we formulate a set of intuitive propagation principles dictating how preferences should be combined, among which the one stating that if a conflict arise, then the more specific context prevails. We then define precise syntactical conditions that expressions respecting the propagation principles should satisfy, and provide a canonical way for doing so. Finally, we consider the well-known Pareto and Prioritized operators and discuss how they implement the proposed framework
Multimedia, Similarity, and Preferences: Adding Flexibility to Your Information Needs
No full textStarting from the 90’s, it was easily recognized that commonly adopted search paradigms were not enough to deal with at-the-time emerging novel DB applications, in which the presence of multimedia data and high dimensionality were both key aspects. In this paper we survey the research activity of our group in the last 25 years, therefore going through issues such as indexing, approximate query processing, and support for preference queries, which are now quite well understood. In doing this we also consider the need to provide the users with simple but powerful tools, able to smooth the processes of query creation/customization and of result interpretation. We complete with a look to the novel issues that the “Big Data” era brings to us
Bounding the cardinality of aggregate views through domain-derived constraints
No full textAccurately estimating the cardinality of aggregate views is crucial for logical and physical design of data warehouses. This paper proposes an approach based on cardinality constraints, derived a-priori from the application domain, which may bound either the cardinality of a view or the ratio between the cardinalities of two views. We face the problem by first computing satisfactory bounds for the cardinality, then by capitalizing on these bounds to determine a good probabilistic estimate for it. In particular, we propose a bounding strategy which achieves an effective trade-off between the tightness of the bounds produced and the computational complexity. © 2002 Elsevier Science B.V. All rights reserved
Flexible Skylines: Dominance for Arbitrary Sets of Monotone Functions
No full textSkyline and ranking queries are two popular, alternative ways of discovering interesting data in large datasets. Skyline queries are simple to specify, as they just return the set of all non-dominated tuples, thereby providing an overall view of potentially interesting results. However, they are not equipped with any means to accommodate user preferences or to control the cardinality of the result set. Ranking queries adopt, instead, a specific scoring function to rank tuples, and can easily control the output size. While specifying a scoring function allows one to give different importance to different attributes by means of, e.g., weight parameters, choosing the “right” weights to use is known to be a hard problem.
In this article, we embrace the skyline approach by introducing an original framework able to capture user preferences by means of constraints on the weights used in a scoring function, which is typically much easier than specifying precise weight values. To this end, we introduce the novel concept of F -dominance, i.e., dominance with respect to a family of scoring functions F : a tuple t is said to F -dominate tuple s when t is always better than or equal to s according to all the functions in F .
Based on F -dominance, we present two flexible skyline (F-skyline) operators, both returning a subset of the skyline: ND, characterizing the set of non-F -dominated tuples; PO, referring to the tuples that are also potentially optimal, i.e., best according to some function in F . While ND and PO coincide and reduce to the traditional skyline when F is the family of all monotone scoring functions, their behaviors differ when subsets thereof are considered. We discuss the formal properties of these new operators, show how to implement them efficiently, and evaluate them on both synthetic and real datasets
Speaking Words of WISDOM: Web Intelligent Search based on DOMain ontologies
No full textIn this paper we present the architecture of a system for searching and querying information sources available on the web which was developed as part of a project called WISDOM. key feature of our proposal is a distributed architecture based on (i) the peer-to-peer paradigm and (ii) the adoption of domainontologies. at the lower level, we support a strong, ontology-based integration of the information content of a bunch of source peers, which form a so-called semantic peer. at the upper level, we provide a loose, mapping-based integrationof a set of semantic peers. we then show how queries can be efficiently managed and distributed in such a two-layer scenario
Flexible Skylines: Dominance for Arbitrary Sets of Monotone Functions
No full textSkyline and ranking queries are two popular, alternative ways of discovering interesting data in large datasets. Skyline queries are simple to specify, as they just return the set of all non-dominated tuples, thereby providing an overall view of potentially interesting results. However, they are not equipped with any means to accommodate user preferences or to control the cardinality of the result set. Ranking queries adopt, instead, a specific scoring function to rank tuples, and can easily control the output size. While specifying a scoring function allows one to give different importance to different attributes by means of, e.g., weight parameters, choosing the "right"weights to use is known to be a hard problem. In this article, we embrace the skyline approach by introducing an original framework able to capture user preferences by means of constraints on the weights used in a scoring function, which is typically much easier than specifying precise weight values. To this end, we introduce the novel concept of F-dominance, i.e., dominance with respect to a family of scoring functions F: A tuple t is said to F-dominate tuple s when t is always better than or equal to s according to all the functions in F. Based on F-dominance, we present two flexible skyline (F-skyline) operators, both returning a subset of the skyline: nd, characterizing the set of non-F-dominated tuples; po, referring to the tuples that are also potentially optimal, i.e., best according to some function in F. While nd and po coincide and reduce to the traditional skyline when F is the family of all monotone scoring functions, their behaviors differ when subsets thereof are considered. We discuss the formal properties of these new operators, show how to implement them efficiently, and evaluate them on both synthetic and real datasets
Getting the Best from Uncertain Data
No full textThe skyline of a relation is the set of tuples that are not dominated by any other tuple in the same relation, where tuple u dominates tuple v if u is no worse than v on all the attributes of interest and strictly better on at least one attribute. Previous attempts to extend skyline queries to probabilistic databases have proposed either a weaker form of domination, which is unsuitable to univocally define the skyline, or a definition that implies algorithms with exponential complexity. In this paper we demonstrate how, given a semantics for linearly ranking probabilistic tuples, the skyline of a probabilistic relation can be univocally defined. Our approach preserves the three fundamental properties of skyline: 1) it equals the union of all top-1 results of monotone scoring functions, 2) it requires no additional parameter to be specified, and 3) it is insensitive to actual attribute scales. We also detail efficient sequential and index-based algorithms
Automatically Joining Pictures to Multiple Taxonomies
No full textAutomatically providing semantics to multimedia objects is still a major open problem. In this paper we describe recent advances within this context and how they have been implemented within the Scenique image retrieval and browsing system. Scenique is based on a multi-dimensional model, where each dimension is a tree-structured taxonomy of concepts, also called semantic tags, that are used to describe the content of images.We describe an original algorithm that, by exploiting low-level visual features, tags, and metadata associated to an image, is able to predict a high-quality set of semantic tags for that image
- …
