Athman Bouguettaya

On-line clustering

We focus on the stability and behavior of three widely used clustering algorithms. Various correlation coefficients are computed to help us understand the sensitivity of object clustering. Surprisingly, the results indicate that there is almost no difference among any clustering approach. Furthermore, all methods appear to be near stable. These findings tend to show that the clustering algorithms are independent of the way objects are inherently clustered.

Ontological Approach for Information Discovery in Internet Databases

The Internet has solved the age-old problem of network connectivity and thus enabling the potential access to, and data sharing among large numbers of databases. However, enabling users to discover useful information requires an adequate metadata infrastructure that must scale with the diversity and dynamism of both users interests and Internet accessible databases. In this paper, we present a model that partitions the information space into a distributed, highly specialized domain ontologies. We also introduce inter-ontology relationships to cater for user-based interests across ontologies defined over Internet databases. We also describe an architecture that implements these two fundamental constructs over Internet databases. The aim of the proposed model and architecture is to eventually facilitate data discovery and sharing for Internet databases.

On the Representation of Objects with Polymorphic Shape and Behaviour

In this paper we present the concept of roles as an extension to object-oriented database models in which identifiable groups of objects are allowed to evolve by assuming transient behavior, while preserving their identity. This mechanism can be used to implement multi-faceted, or polymorphic, objects by allowing the same object to be represented under different perspectives — possibly at discrete points in time — with each perspective having its own state and behavior. The concept of roles is used to implement dynamic properties and aspects of objects and provides for added modeling power and flexibility.

Large Multidatabases: Issues and Directions

Abstract The level of complexity in achieving interoperability in a multidatabase system largely depends on the number of component databases. If the size of the multidatabase system is small, it is reasonable to assume that current technology is adequate to address the interoperability problem. However, as the size increases, it becomes more and more difficult (if not impossible) to apply existing technology. The size of the multidatabase system is the single most important criterion that shapes the complexity of the problem. As the size becomes larger, new problems related to database autonomy and heterogeneity, as well as user education about available information, surface.

Data clustering analysis in a multidimensional space

Abstract Cluster analysis techniques are used to classify objects into groups based on their similarities. There is a wide choice of methods with different requirements in computer resources. We present the result of a fairly exhaustive study to evaluate three commonly used clustering algorithms, namely, single linkage, complete linkage, and centroid. The cluster analysis study is conducted in the two dimensional (2-D) space. Three types of statistical distribution are used. Two different types of distances to compare lists of objects are also used. The results point to some startling similarities in the behavior and stability of all clustering methods.

Incremental specification of views across databases

In an environment consisting of a large number of databases, it can no longer be reasonable to assume that construction of users views could rely on an a priori static partial or global schema integration. Approaches which are based on integrated mechanisms are obviously intractable (if not impossible) within a large size environment. The authors propose an approach where users are allowed to dynamically investigate information and incrementally specify their views. The idea is to use a functional approach to model the construction of views. Further, their approach provides a mechanism by which views adapt to the evolution of the information space. This mechanism relies on dynamic triggers to fulfil this goal.<<ETX>>

Using Java and CORBA for implementing Internet databases

We describe an architecture called WebFINDIT that allows dynamic couplings of Web accessible databases based on their content and interest. We propose an implementation using WWW, Java, JDBC, and CORBAs ORBs that communicate via the CORBAs IIOP protocol. The combination of these technologies offers a compelling middleware infrastructure to implement fluid-area enterprise applications. In addition to a discussion of WebFINDITs core concepts and implementation architecture, we also discuss an experience of exiting WebFINDIT in a healthcare application.

On building a hyperdistributed database

Sharing data among disparate databases has so far mostly been achieved through some form of ad-hoc schema integration. This approach becomes less tractable as the number of participating database increases. Therefore, the complexity of making autonomous heterogeneous databases interoperate is dependent on adequately addressing the autonomy and heterogeneity issues. In this paper, we describe a prototype that implements an approach which addresses these issues in the context of large multidatabase systems. In particular, we describe a scheme that builds a Hyperdistributed Database using a two-staged approach. We also describe how conglomerations of databases are formed, modified, and evolved.

World Wide Database—integrating the Web, CORBA and databases

1 Introduction While most of the data published on the Web is either semi-structured (e.g., HTML documents) or unstructured (e.g., text files, images), the Web also offers hooks to ac.cess non-Web centric structured data (e.g., relational databases). CGI scripts are usually used to access back-end databases. The Web has so far been incongruous with databases. The reason that the Web is database unfriendly is that it has been developed for open data sources. Databases are closed in nature in that communication with them is through a rigid protocol (DBMS). One needs to know the schema of a database to access or modify its state. This is fundamentally different from the openness and freeform type of Web data. 1Veb protocols and search engines have been developed for this kind of requirements and environments. Therefore, it is important to note that information retrieval and search techniques could not be applied because of the different nature of, and the fundamental assumptions about the data. However, we could not discount the fact that the Web has made it now possible to have one single interface to potentially access all Internet accessible databases. The challenge now is to make the Web database friendly. This essentially means that we need to build an adequate infrastructure on top of the Web that will provide for a World Wide Database (WWD). In order to address problems of scalability and language support for Web accessible databases, the WebFINDIT prototype has been developed [3, 11. WebFINDIT is a system for describing, locating and accessing data in Internet-accessible databases. We present an architecture and supporting tools that enable users to bu:ild complex and emerging Web applications in a simple and flexible way. The main idea is to incorporate simple access to Web-accessible databases as well as support for scalability and extensibility into a flexible interoperable architecture for large information spaces. Most existing techniques focused extensively on either data sharing among small numbler of heterogeneous databases or on information discovery and brokering in the context of unstructured or semi-structured Web-resident data [l]. In our research, we consider issues to achieve the database equivalent of the WWW. We call it the World Wide Database (WWD). The WWD will be fundamentally different from the WWW in that databases are relatively closed as compared to traditional Web-resident data. Therefore, we need to devise an infrastructure that will recognize the specificities of accessing databases …

FINDIT: a server based approach to finding information in large scale heterogeneous databases

Finding information in large scale autonomous heterogeneous databases is an issue that has been virtually unexplored in database research. The authors describe a new approach for finding information in a large scale network of autonomous and heterogeneous databases. The system relies on an object-oriented architecture to educate users about the available space of information and help them locate information sources. The system, named FINDIT, accommodates database autonomy as well as heterogeneity. To ensure information abstraction and encapsulation the system is divided into three categories, namely, user servers, database servers, and databases. User servers are geared towards satisfying users needs while database servers are in charge of managing databases.<<ETX>>