Wyniki wyszukiwania - BazTech

1

Abduction for Accessing Information Sources

Meghini C., Tzitzikas Y., Spyratos N.

Fundamenta Informaticae

|

2008

|

Vol. 83, nr 4

355-387

EN

We consider a general form of information sources, consisting of a set of objects classified by terms arranged in a taxonomy. The query-based access to the information stored in sources of this kind, is plagued with uncertainty, due, among other things, to the possible linguistic mismatch between the user and the object classification. To overcome this uncertainty in all situations in which the user is not finding the desired information and is not willing or able to state a new query, the study proposes to extend the classification, in a way that is as reasonable as possible with respect to the original one. By equating reasonableness with logical implication, the sought extension turns out to be an explanation of the classification, captured by abduction. The problem of query evaluation on information sources extended in this way is studied and a polynomial time algorithm is provided for the general case, in which no hypothesis is made on the structure of the taxonomy. The algorithm is successively specialized on a most common kind of information sources, namely sources whose taxonomy can be represented as a directed acyclic graph. It is shown that query evaluation on extended sources is easier for this kind of sources. Finally, two applications of the method are presented, which capture very important aspects of information access: information browsing and query result ranking.

2

Deriving and retrieving contextual categorical information through instance inheritance

Analyti A., Spyratos N., Constantopoulos P.

Fundamenta Informaticae

|

2000

|

Vol. 44, Nr 4

321-351

EN

In semantic and object-oriented data models, each class has one or more typing properties that associate it to other classes, and carry type information about all instances of the class. We introduce a new kind of property that we call instance-typing property. An instance-typing property associates an instance of a class to another class, and carries type information about that particular instance (and not about all instances of the class). Instance-typing properties are important as they allow to represent summary information about an instance, in addition to specific information. In this paper, we study inheritance of properties from a class to an instance, using type information about the class, as well as type information about the instance. This kind of inheritance, that we call contextual instance-inheritance, provides us with the most specific type information about the instance, in a particular context. Intuitively, a context is a metaclass of interest with respect to which this most specific information is determined. We demonstrate that contextual instance-inheritance is a powerful conceptual modeling mechanism, capable of expressing valuable information about instances. We also provide a framework in which derived instance-inherited properties can be represented and retrieved in the same way as "usual" properties.

3

On the semantics of a semantic network

Analyti A., Spyratos N., Constantopoulos P.

Fundamenta Informaticae

|

1998

|

Vol. 36, nr 2,3

109-144

EN

We elaborate on the semantics of an enhanced object-oriented semantic network, where multiple instantiation, multiple specialization, and meta-classes are supported for both kinds of objects: entities and properties. By semantics of a semantic network, we mean the information (both explicit and derived) that the semantic network carries. Several data models use semantic networks to organize information. However, many of these models do not have a formalism defining what the semantics of the semantic network is. In our data model, in addition to the Isa relation, we consider a stronger from of specialization for properties, that we call restriction Isa, or Risa for short. The Risa relation expresses property value refinement. A distinctive feature of our data model is that it supports the interaction between Isa and Risa relations. The combination of Isa and Risa provides a powerful conceptual modeling mechanism. The user declares objects and relations between objects through a program. Reasoning is done through a number of (built-in) inference rules that allow for derivations both at instance and schema level. Through the inference rules, new objects and new relations between objects are derived. In our data model, inherited properties are considered to be derived objects. In addition to the inference rules, a number of (built-in) system constraints exist for checking the validity of a program.