Showing papers presented at "Flexible Query Answering Systems in 2002"

Bipolarity in Flexible Querying

Abstract: The paper advocates the interest of distinguishing between negative and positive preferences in the processing of flexible queries. Negative preferences express what is (more or less, or completely) impossible or undesirable, and by complementation state flexible constraints restricting the possible or acceptable values. Positive preferences are not compulsory, but rather express wishes; they state what attribute values would be really satisfactory. The paper discusses the handling of bipolar queries, i.e. queries involving negative and positive preferences, in the framework of possibility theory. Both ordinary queries expressed in terms of requirements, and case-based queries referring to examples, are considered in this perspective.

Consistent Answers from Integrated Data Sources

Abstract: When data sources are integrated into a single global system, inconsistencies wrt global integrity constraints are likely to occur. In this paper, the notion of consistent answer to a global query in the context of the local-as-view approach to data integration is characterized. Furthermore, a methodology for generating query plans for retrieving consistent answers to global queries is introduced. For this purpose, an extension of the inverse-rules algorithm for derivingquery plans for answering first-order queries is presented.

An Augmented Visual Query Mechanism for Finding Patterns in Time Series Data

Abstract: Relatively few query tools exist for data exploration and pattern identification in time series data sets. In previous work we introduced Time-boxes. Timeboxes are rectangular, direct-manipulation queries for studying time-series datasets. We demonstrated how Timeboxes can be used to support interactive exploration via dynamic queries, along with overviews of query results and drag-and-drop support for query-by-example. In this paper, we extend our work by introducing Variable Time Timeboxes (VTT). VTTs are a natural generalization of Timeboxes, which permit the specification of queries that allow a degree of uncertainty in the time axis. We carefully motivate the need for these more expressive queries, and demonstrate the utility of our approach on several data sets.

Approximating Terminological Queries

Abstract: Current proposals for languages to encode terminological knowledge in intelligent systems support logical reasoning for answering user queries about objects and classes. An application ofthese languages on the World Wide Web, however, is hampered by the limitations of logical reasoning in terms ofefficiency and flexibility. In this paper we describe, how techniques from approximate reasoning can be used to overcome these problems. We discuss terminological knowledge and approximate reasoning in general and show the benefits of approximate reasoning using the example of building and maintaining semantic catalogues that can be used to query resource locations based on object classes.

Spatio-Temporal Querying in Video Databases

Abstract: In this paper a video data model that allows efficient and effective representation and querying of spatio-temporal properties of objects is presented. The data model is focused on the semantic content of video streams. Objects, events, activities performed by objects are main interests of the model. The model supports fuzzy spatial queries including querying spatial relationships between objects and querying the trajectories of objects. The model is flexible enough to define new spatial relationship types between objects without changing the basic data model. A prototype of the proposed model has been implemented. The prototype allows various spatio-temporal queries along with the fuzzy ones and it is prone to implement compound queries without major changes in the data model.

On Measuring Similarity for Conceptual Querying

Abstract: The focus of this paper is approaches to measuring similarity for application in connection with query evaluation. Rather than only comparing at the level of words the issue here is to compare concepts that appear as compound expressions derived from list of words through brief natural language analysis. Concepts refers to and are compared with respect to an ontology describing the domain of the database. We discuss three different principles for measuring similarity between concepts. One in the form of subsumption expansion of concepts and two as different measures of distance in a graphical representation of an ontology.

X-Compass: An XML Agent for Supporting User Navigation on the Web

Abstract: In this paper we present X-Compass, an XML agent for supporting a user during her/his navigation on the Web. This agent is the result of our attempt of synthesizing, in a unique context, important guidelines currently characterizing the research in various Computer Science sectors. X-Compass constructs and handles a rather rich, even if light, user profile. This latter is, then, exploited for supporting the user in the efficient search of information of her/his interest; in this way, the proposed agent behaves as a content-based recommender system. Moreover, X-Compass is particularly suited for constructing multi-agent systems and, therefore, for implementing collaborative filtering recommendation techniques. In addition, being based on XML, X-Compass is particularly light and capable of operating on various hardware and software platforms. Finally, the exploitation of XML makes the information exchange among X-Compass agents and, therefore, the management and the exploitation of X-Compass multi-agent systems, easy.

Adding Flexibility to Structure Similarity Queries on XML Data

Abstract: The presence of structure inside XML documents poses the hard challenge of providing flexible query matching methods for effective retrieval of results In this paper we present an approach that faces this issue in a twofold fashion: 1) it exploits new approximations on data structure; 2) it provides a relevance ranking method that takes into account the degree of correctness and completeness of results with respect to a given query, as well as the degree of cohesion of data retrieved

Theory of K-Calculuses as a Powerful and Flexible Mathematical Framework for Building Ontologies and Designing Natural Language Processing Systems

Abstract: The necessity of developing more powerful and convenient formal means (in comparison with the widely used ones) for describing structured meanings of natural language (NL) texts and building ontologies is grounded. The basic principles of an original mathematical approach to this problem are outlined; this approach is given by the the theory of K-calculuses and K-languages (the KCL-theory) elaborated by the author. The considered examples of building semantic representations of the NL-texts, of constructing definitions of concepts, and of finding an answer to a question of the end user pertain to biology, medicine, ecology, and business. The advantages of the KCL-theory in comparison with Discourse Representation Theory, Theory of Conceptual Graphs, and Episodic Logic are stated. It is reported about a number of successful applications of the KCL-theory to the design of NL-processing systems. The metaphor "a kitchen combine" is used for underlining that the KCL-theory provides a broad spectrum of new opportunities for the design of NL-processing systems and building ontologies.

Interactive Query Formulation in Semistructured Databases

Abstract: The use of large amounts of distributed and heterogeneous information has become extremely cumbersome; this difficulty is mainly related to exploring the data, rather than actually storing or exchanging it. The user who is interested in small bits of information is getting more and more confused when having to dig under a large volume of diverse and more importantly semi-structured data. In this paper, we propose an interactive and adaptive framework that guides the user in the search for data, by disclosing only a part of the underlying information at a time. It first provides the user with a high-level view of the raw data and gradually adapts to his/her needs in order to offer a refined answer. The proposed model offers the possibility to query a semistructured database based on general schema-related constraints imposed by the user or identified by the system, but without specific knowledge of the underlying metadata. This is achieved by receiving initially an amorphic query, which may consist of one or more basic paths, and helping the user to refine it gradually to a specific semistructured query, expressed in a language like XQuery or Lorel.

Conjunctive Aggregation of Extended Possibilistic Truth Values and Flexible Database Querying

Abstract: Extended possibilistic truth values are a flexible means to model the linguistic (un)certainty about the truth value of a proposition. With respect to flexible database querying, they are suited to express the extent to which a database instance satisfies a query condition. Since a query can impose several conditions, aggregation is necessary. In this paper, three definitions of conjunctive aggregation operators for extended possibilistic truth values are presented and compared with each other on the basis of their capability to rank the possible alternatives in the result of a (flexible) database query.

Paraconsistent Query Answering Systems

Abstract: Classical logic predicts that everything (thus nothing useful at all) follows from inconsistency. A paraconsistent logic is a logic where inconsistency does not lead to such an explosion, and since in practice consistency is difficult to achieve there are many potential applications of paraconsistent logics in query answering systems. We compare the paraconsistent and the non-monotonic solutions to the problem of contradictions. We propose a many-valued paraconsistent logic based on a simple notion of indeterminacy. In particular we describe the semantics of the logic using key equalities for the logical operators. We relate our approach to works on bilattices. We also discuss and provide formalizations of two case studies, notably the well-known example involving penguins and a more interesting example in the domain of medicine.

Query Construction through Meaningful Suggestions of Terms

Abstract: Query formulation by using database specific query languages such as SQL or OQL turns out to be cumbersome or even impossible when end-users need to pose queries to large database schemes. This is due to the difficulties which arise out of the wrong or impossible interpretations of storage models and the lack of mechanisms to embed application domain semantics within query languages. Visual query languages (VQLs) and natural language (NL) based query interfaces in query answering systems alleviate, in some cases, the task of constructing a query. VQLs, however, are bound to visual formalisms which need further interpretation and still lack the use of semantics other than those provided by well-known conceptual models (EER, OMT, etc.). NL based approaches, on the other side, presuppose a considerable knowledge of the vocabulary terms to be used by the end-user for a particular application domain and, furthermore, they do not exploit the meaning of words other than that as provided by the syntax, in order to formulate a meaningful query. This turns out to be cumbersome, especially when advanced terminologies and large vocabularies should be used. This is also strengthened by the non-unique name assumption characterizing the application domain vocabulary. In this paper, we present a query construction paradigm which underlies the Meaning Driven Data Query Language MDDQL. It strongly relies on the construction of queries through suggestions of meaningful terms, in a preferred natural language, when requested by the end-user. An inference engine is responsible for the suggestion of a semantically consistent set, not only of application domain terms, but also of operator or operation terms, during the query construction process. All inferences are drawn at a "heuristic level" of representation of the vocabulary, i.e., on the basis of data structures (cyclic graph), and not at an "epistemological level", i.e., based on logic-like representations.

A Parallel Spatial Join Processing for Distributed Spatial Databases

Abstract: In recent years, there have been needs of accessing spatial data from distributed and preexisting spatial database systems interconnected through a network. In a distributed environment, spatial joins for two spatial relations residing at geographically separated sites are expensive in terms of computation and transmission cost because of the large size and complexity of spatial data. Previous distributed algorithm based on the spatial semijoin has accomplished performance improvements by eliminating objects before transmission to reduce both transmission and local processing costs. But with a widespread of a high bandwidth data transmission, the parallelism through data redistribution may improve the performance of spatial joins in spite of additional transmission costs. Hence, we propose a parallel spatial join processing for distributed spatial databases. We apply the task distribution method minimizing the data transmission and the solution for task distribution using a graph partitioning method. In experiments, we showed that the proposed method provides useful reductions in the cost of evaluating a join.

Extending Relational Data Access Programming Libraries for Fuzziness: The fJDBC Framework

Abstract: Fuzzy relational databases have been extensively studied in recent years, resulting in several models and representation techniques, some of which have been implemented as software layers on top of diverse existing database systems. Fuzzy extensions to query languages and end-user query interfaces have also been developed, but the design of programming interfaces has not been properly addressed yet. In this paper, we describe a software framework called fJDBC that extends the Java Database Connectivity API by enabling fuzzy queries on existing relational databases, using externally-stored metadata. Since the main design objective of this extension is usability for existing database programmers, only a restricted subset of extensions (supported also by an extended object modelling notation) has been included. The overall design of the framework and some empirical results are also described.

Association Rule Extraction for Text Mining

Abstract: We present the definition of fuzzy association rules and fuzzy transactions in a text framework. The traditional mining techniques are applied to documents to extract rules. The fuzzy framework allows us to deal with a fuzzy extended Boolean model. Text mining with fuzzy association rules is applied to one of the classical problems in Information Retrieval: query refinement. The extracted rules help users to query the system by showing them a list of candidate terms to refine the query. Different procedures to apply these rules in an automatic and semi-automatic way are also presented.

A Knowledge-Based Query System for Biological Databases

Abstract: In more than one decade, genomic research produced a huge amount of experimental data. Although these data are usually freely available on the World Wide Web, accessing them in a consistent and fruitful way is not always an easy task. One of the main causes of this problem can be recognized in the lack of user interfaces that are sufficiently flexible and, at the same time, highly interactive and cooperative and also semantically precise.This paper tackles this issues by presenting a semantics drive paradigm for formulating queries to genomic and protein databases. The paradigm is founded on knowledge-based reasoning capabilities, in order to provide the user with a semantics driven guide in the difficult task of building the intended query.

Context - Sensitive Query Expansion Based on Fuzzy Clustering of Index Terms

Abstract: Modern Information Retrieval Systems match the terms contained in a user's query with available documentsthrough the use of an index. In this work, we propose a method for expanding the query with its associated terms, in order to increase the system recall. The proposed method is based on a novel fuzzy clustering of the index terms, using their common occurrence in documents as clustering criterion. The clusters which are relevant to the terms of the query form the query context. The terms of the clusters that belong to the context are used to expand the query. Clusters participate in the expansion according to their degree of relevance to the query. Precision of the result is thus improved. This statistical approach for query expansion is useful when no a priori semantic knowledge is available.

User Preference Information in Query Answering

Abstract: This paper discusses the use of usage information to enhance cooperative behaviour from query answering systems. A user can pose a query to the system by providing (at query time) update and preference information. Updates allow us to model dynamically evolving worlds and preferences allow us to facilitate the retrieval of information by targeting the answers of the system with respect to users' interests in a given context.

A Class-Based Logic Language for Ontologies

Abstract: We describe a formal design of a hybrid language and show that it provides a suitable basis for representing and interacting with ontologies. We specify ontologies and query them with a clear logical semantics. The proposed language combines the expressive power of description logics and rules.

Recursive Queries in Product Databases

Abstract: The relational data model does not offer query language solutions for recursive applications. A missing semantic update operation is the fundamental reason for this failure. It is shown that a semantic data model can solve these applications using simple linear constructs for definition and manipulation, without explicit recursion, nesting, iteration or navigation. The new implementation technique is reliable and efficient for application in end user environments. It is based on semantic query specification and metadata processing. Experiments with the semantic Xplain DBMS have confirmed the linear time complexity for product database applications.

Querying Multidatabase Systems Using SIQL

Abstract: Several approaches have been proposed to provide access to data which is spread across different sources. Prominent among these are the approaches taken in federated systems and multidatabase systems. The main thrust is to provide the user with a mechanism to access several databases in a unified way. However, in all these systems the user is not isolated from specifying the path expressions in the query. In our system, we propose a query language, Structure Independent Query Language (SIQL), which has no reference to the structure of the databases. The structure is arrived at by the system by using the component databases that are registered with it. In general, there can be multiple queries that can be generated for a user query. We define 'closeness' criterion which imposes a ordering on the queries that can be generated. The queries are generated and executed according to this ordering and the result is presented to the user.

Intelligent Information Integration as a Constraint Handling Problem

Abstract: Intelligent Information Integration (I3) aims at combining heterogeneous and distributed information sources by explicitly representing and reasoning about their content, giving the user the illusion of interacting with a uniform system. In this paper we show how query planning in such a system can be reduced to a constraint handling problem. Conceptually, our approach relies on a generalized abductive reasoning mechanism involving so called partially open predicates, which support a seamless combination of backward (goal-directed) and forward reasoning. The original aspects of our approach consist in the early detection of (plan) inconsistencies using forward propagation of constraints as well as in the seamless interleaving of query planning and execution. Unlike other specialized query planning algorithms, for which domain reasoning and query planning are only loosely coupled, our encoding of source and domain models into Constraint Handling Rules allows us to fully and efficiently exploit existing domain knowledge. The ability to automatically derive source interactions from domain models (ontologies) enhances the flexibility of modeling.

Representation & Querying of Temporal Conflict

Abstract: In this paper we are concerned with the problem of designing of a conceptual framework for the representation needs of conflicting but still certain data, in multi-source temporal environments We put forward a proposal based on the assumption that conflict and uncertainty are dual properties We believe that the querying of certain but conflicting information bases will always provide possible but not always definite answers We start by positioning conceptual schema, database with regard to this problem as part of a three-layer architecture for capturing and representing the dual properties conflict and uncertainty Then we distinguish lattice structures as the central task in query answering considering the conceptual, instance level always with respect to the dynamism-dimension of data

Flexible Management of Consistency and Availability of Networked Data Replications

Abstract: We describe a family of three replication protocols, each of which can operate in three different modes of consistency. The protocols are tailored to satisfy the availability demands of interconnected databases that have a high degree of data locality. The protocols accomplish a grade of transaction completion which does not compromise availability, and ensure the consistency of replicas also if a transaction needs to be aborted. Flexibility of query answering is understood as optimizing the tradeoff between consistency and availability, i.e., between correctness and timeliness of query answering. This is achieved by choosing an appropriate protocol alternative, and changing the consistency mode of operation during the session, as appropriate for a given transaction.

A Similarity-Based Unification Model for Flexible Querying

Abstract: We use the formal model for similarity-based fuzzy unification in multi-adjoint logic programs to provide new tools for flexible querying. Our approach is based on a general framework for logic programming, which gives a formal model of fuzzy logic programming extended by fuzzy similarities and axioms of first-order logic with equality. As a source of similarities we consider different approaches, such as statistical generation of fuzzy similarities, or similarities generated by some information retrieval techniques or similarities arisingfrom fuzzy conceptual lattices.

A Formal Model for Data Fusion

Abstract: In information retrieval, the data fusion problem is as follows: given two or more independent retrieved sets of ranked documents in response to the same query, how to merge the sets in order to present the user with the most effective ranking? We propose a formal model for data fusion that is based on the knowledge that can be derived from the retrieved documents The modelis based on evidential reasoning, a theory that formally expresses knowledge and the combination of knowledge Knowledge characterising a ranked list of retrieved documents is symbolised The combination of knowledge associated to the several retrieval results yields the characterisation of the merged result

An Efficient Incremental Lower Bound Approach for Solving Approximate Nearest-Neighbor Problem of Complex Vague Queries

Abstract: In this paper, we define a complex vague query as a multifeature nearest neighbor query To answer such queries, the system must search on some feature spaces individually and then combine the results in order to find the final answers The feature spaces are usually multidimensional and may consist of the sheer volume of data Therefore searching costs are prohibitively expensive for complex vague queries For only such a single-feature space, to alleviate the costs, problem of answering nearest neighbor and approximate nearest neighbor queries has been extensively studied and quite well addressed in the literature This paper, however, introduces an approach for finding (1+?)- approximate nearest neighbors of complex vague queries, which must deal with the problem on multiple feature spaces This approach is based on a novel, efficient and general algorithm called ISA-Incremental hyper-Sphere Approach [12, 13], which has recently been introduced for solving nearest neighbor problem in the VQS-Vague Query System [22] To the best of our knowledge, the work presented in this paper is one of the vanguard solutions for generally dealing with problem of approximate multi-feature nearest neighbor queries The experimental results will prove the efficiency of the proposed approach

Effect Preservation as a Means for Achieving Update Consistency

Abstract: Management of data systems must cope with changes, initiated by users or applications. Query answering in a frequently modified data system must be consistent with the updates. A natural expectation is that a fact that was successfully added, is retrievable as long as it was not intentionally removed. Active Databases do not meet this expectation, since contradicting rules may undo events that triggered them. In this paper, we associate database transactions with effects, and present a method that takes care of preserving the effects of updates. We introduce a compile-time effect preservation transformation that revises a transaction so to prevent contradictory updates, i.e., yields an effect preserving transaction. Our method yields an expressive and efficient transaction, since it is based on interleaving run-time sensitive analysis of effects within the compile-time transformation of a transaction. The interleaving and the compile-time reduction of effects account for the efficiency; the run-time sensitivity of effects accounts for the expressiveness. In the context of Active Databases this method can be used to statically revise a rule application plan, so to prevent contradictory updates. We claim that an effect preserving method should be integrated into every update and query processing system.